Datasets:

64bits

/

lex_fridman_podcast_for_llm_vicuna

like 14

[ { "from": "human", "value": "You know, it's my- Wait a minute, wait a minute they're not, listen. No, as I did, I'm not even touching that one. It is true, it is true. The document, there is a-" }, { "from": "gpt", "value": "There is a document, it is true. There is a documentary where they duked a bunch of scientists who are like oceanographers and they like showed them ancient footage of, you know, mariners saying that seals were mermaids. Who cares? It's, I was young, I got brought to Hollywood and I got spit out the other side, and that's on me. That's not their fault. You know, there's that parable about the frog who gives the scorpion a ride across the water and then at the end he says, I'll give you a ride just don't sting me and then I get to the other side and the scorpion stings them and the frog goes, why did you do that? And the scorpion goes, I'm a scorpion." }, { "from": "human", "value": "It's not their fault. That's my nature but now that you've become much more well known and much more successful at what you do, you have a platform. Can you return to those people and use it, the machine, to get more and more attention. Is that something you work on?" }, { "from": "gpt", "value": "Or do you prefer to work completely outside? I think that most of the success that we've had now in protecting the rainforest and the levels that we've reached are so far. I think back to those barefoot days of catching snakes with JJ on the boat and now the massive ecological reserve we have and the team of rangers and the converted loggers and all of that is because of the ability to communicate and to show people, but that's all been through social media. And so I'm open to the fact, if somebody came and gave a sort of like Bourdainian pass where they said, look, you can be yourself, you can swear, you can fart, you can smoke, you can do whatever you wanna do, go out there and show us the real thing, I would love to. But now I know how those contracts need to be. I need to have right to refusal and they can't change them. And so I'd almost rather just do it like the way I think like Mr. Beast does stuff where it's like you just, you get a crew of guys and some seed money and go film the episodes and put it out exactly, I mean, a committee never helped real art be better. It has to come from the source of inspiration. So you get, I think, you know, you get JJ and a crew of people or the guys in Africa that I'm working with right now doing elephant conservation and like, but you gotta show real. I mean, look, that's why, I mean, look, that's why Joe Rogan is important right now. That's why you're important right now is because it's not being filtered through this ridiculous system of polishing it" }, { "from": "human", "value": "and dumbing it down. Yeah, that's why Joe has been an inspiration. You don't need a crew of a bunch of people. You don't need a crew period. All you need is one or two other people and that's it. And in my case, you don't need anybody. I've been doing this by requirement. I just need to be by myself. There's a few other folks now that are helpful with the editing and so on, but it's just, they make life more awesome as opposed to a boss that tells, a creative director. Somebody told me, actually, I was visiting LA, I think it was in LA, where they were saying that now for all intimate scenes in Hollywood movies, there's an intimacy director. So when there's two people having sex, there's a third person that ensures that, on film, so it's not real, but there's still intimacy, there's a third person that ensures that everyone is comfortable and the actors say that this always ruins the chemistry of the scene. And so, it's a very Hollywood thing. I understand there's creepy people and so on. Yeah, I understand. Thanks, Harvey Weinstein. It's usually, I think, comes from the director pushing things too hard. If you just leave it to the actors, they know their boundaries, they can control their own boundaries. So the intimacy director is more for the director pushing things." }, { "from": "gpt", "value": "There's Werner Herzog, you know, there's Werner Herzog. I understand the logic. Let's make sure that we don't have anything happen here that shouldn't be happening. I get it, but I, yeah. But no, I think that authenticity is the greatest currency. And I think that in order for me to tell the stories that I can tell, like, you know, what changed the game for me was, I wanna tell you this story. So in 2019, the Amazon fire started popping off and we had just gone to film like a month earlier. We'd filmed like a small documentary and they'd been following me as if I was on a solo, which, you know, we did the best we could. I lived on my own. But as we were driving, we passed a spot where the flames were 70 feet tall. The forest was being destroyed. And I went out there with my phone, which overheated in like two minutes and said, you can't use it. But for a second, I was out there in the flames, picking up animals and throwing them off to try and just get them cooled off. I was trying to get snakes out of there. Everything was, the birds are flying, and I fucking lost it. I was red-eyed, I was crying, and I was going, this is happening every fucking day, I was screaming, and it's the first time that I'd done that. Because I've seen the burning so many times and I just lost it that day. And I don't know what made me pull out my phone because, usually, in those intense moments I say, forget the documentation, this is real life. we've got stuff to do and I'm doing, I'm not documenting. And then a month later, I'm home and I'm in New York and all of a sudden I see these articles like, the Amazon's burning worse this year than it was last year and blah, blah, blah. And I was like this, you know, fucking. And I threw it up on Instagram like eight o'clock at night and I'd never, like I'd never cursed on Instagram. I don't know why, I just never did. And my phone was on 100% and I put it on top of the refrigerator and I went to bed and I woke up in the morning and my phone was on the floor on 2% and it was ringing off the hook. And it was like the news. And they were are you the guy that posted that viral video about the Amazon? And I was like, what? And that was the start, that's where it broke. And that's where we went from barefoot in the Amazon to all of a sudden, I was talking head for three weeks and going around on all these news stations and all of a sudden I was like the spokesperson for the Amazon and JJ's calling me and he was like, go, go, go, go, go, like get us that support. and it was just, you know, so communicating with people and bringing them into that reality and whether it's, you know, rhino poaching and elephant poaching or the Amazon being destroyed. It's like, to me, it's like being able to take people" } ]

[ { "from": "human", "value": "into that is something that I would love to do. Yeah, and you do it directly with authenticity on your Instagram. People should definitely follow your Instagram. I think, I think Rogan follows your Instagram too." }, { "from": "gpt", "value": "Well, the end of that story actually kind of involves him because yeah, because I went to all these news outlets and I was living in green rooms and traveling around and I was all strung out and I hadn't seen anybody I actually know in a few weeks, which was starting to get to me. And I finally got home and I went to a family party and everyone was like, dude, you've been, it's been crazy, and I was like, yup. And then I left and my cousin, Michael calls me and he's screaming. And I'm going, what, what, what, what, what? And he goes, Joe Rogan just shared it, Joe Rogan just shared it. And everyone was losing their shit and it was so amazing. And it was like, that's when it really took off. And what happened as a result of all of this is that a Canadian entrepreneur who started light speed reached out and several months later after COVID, after that boom, I'd been in the game for maybe 13 years or something. Had no money, no savings, no job, no nothing. And after that great publicity thing, nothing happened. The waves came and everything got real exciting. Everybody reached out and they said, we care so much. Nothing happened though. We can run into battle, but if we don't have arrows in the quiver, what can we do? And I actually, I made a phone call to my friend Mohsen right at the start of COVID and I was going through divorce and I was broke. And I said, I'm gonna get a job. I said, I give up. So this is stupid. I said, the equal tourism business is done. Jungle keepers has dried up. We're done. And then this guy Dax De Silva called me on the phone and said, listen, I'm in, what do we gotta do? And so if the analogy was me and JJ and a few other people trying to hold this boulder back from just destroying the rainforest, all of a sudden Dax comes in like a Titan and just puts his arm out and just goes, I'm gonna help. And he gave us the funding to start actually developing a ranger program, to start actually bringing loggers, to be protectors of the forest, to be supporting smaller conservation things. And now we're protecting 50,000 acres of rainforest. We're protecting entire streams and ecosystems that I love. And we're soon gonna double that. And it's like this whole thing. So yeah, the communication of these things is crucial. And I actually think it's incredible" }, { "from": "human", "value": "that social media has played such a big role in it. Well, I mean, just cause I know Joe well and I love him so much, I definitely think you should do his podcast, but also just be friends with him. I think you guys, he's one, not the meme, but he's one with nature and not much more with the, I'm one, while I do appreciate and love nature, I also love technology and robots and so on." }, { "from": "gpt", "value": "So we're in that meme type of way, we're very different. Well, either way, at some point, make sure you tell the guy, thank you, because it definitely really helped push us over that limit where, if enough people see it, you get someone like Dax who says, I can help and I have the resources to help. And that changed our whole lives." }, { "from": "human", "value": "Back to the jungle, you had a bunch of interactions with jaguars. How are you still alive?" }, { "from": "gpt", "value": "Man, dude, Jags aren't the, Jags, I'll tell you this, Jags are not the danger, the falling trees are the danger. I'll tell you some elephant stories and then you'll wonder why I'm still alive. But Jags, I've just one, you know, so JJ started in Santiago, his dad started challenging me to do solos. Go out alone into the wild style. You know, I'd have a hammock, a headlamp, three days worth of food, some fish hooks, a machete. That's it. And so like one of the stories that happened early on was I was out there and it was raining and I was lost. And this is how we test your jungle knowledge. Can you survive out there? Do you know how to find food? Have you listened to the things that we taught you? And there was one night that I was in a hammock and a Jaguar came up and I was asleep when it happened and she came up right next to my head and she was and I could hear her smelling me. And then my first instinct was to turn on my headlamp and just the sound of my arm moving against the material. And she just, she just right here I could feel her breath and I just laid there in the dark. And that's one of those moments where you go, you really learn a lot about yourself, because I wasn't scared. I felt like I understood the intentions of the cat. If she was hunting, I'd already be dead. She was curious, and I was lost, and I didn't know if I was ever gonna get out of that jungle, but what she did was energize me, because it was an experience like the giant anaconda, where I said, this is so wild, that it's so almost cinematically outside the realm of what I thought my life could be like, that it made me like, wait, because the previous day, I was lost, tired, confused, kind of devastated, tail between my legs. After that, I was like, man, you've been waiting for this your whole life. Go get it. And I woke up, and I was like, I am gonna navigate. Even though I've been in this swamp for three days, I'm gonna find my way out of this swamp, and she just breathed fire into me, where it was like, if that's possible, if I could be six inches away from a jaguar's face," } ]

[ { "from": "human", "value": "then I got that energy from her. So that you were able to start to really hear and feel the jungle around you? Yeah. That was a sign that you know what you're doing. It really felt like a sign, I really did. How do you survive in a solo hike to the jungle? What are the different components, what are the different dangers? So you said you had a hammock, you had some food." }, { "from": "gpt", "value": "What kind of food, by the way, we're talking? Yeah. So you said you had a hammock. Nuts, stuff that won't go bad, because you can't really start a fire in the Amazon. Like, I'm a good fire, I mean, I camp all over the place. I'm a wilderness guide. Starting a fire in the Amazon is futile. In fact, a lot of survival manuals will tell you, don't do it, because if you're really lost, it'll break your spirit. You're not gonna be able to do it. That's dark. Yeah, they're like, don't even try it. But you can still get like hyperthermia from, if you get wet and you lay out in the jungle, you can still, exposure can still get you, so you want fire. Yeah. I even, in the beginning I used to bring like ramen noodles, which is, nutritionally is irrelevant. So I started bringing like nuts, and then supplementing that with fish, which forced me to become a very good fisherman. Now, of course, JJ knows that he can like, they can cut certain roots, and they bash it up, and they put it in the stream, and the fish just float to the top, and they take with, so he's got all the cheat codes, whereas, I'm sitting there with a hook, and then he's like he'll go, now find bait.\" And I go, bait. In the most competitive ecosystem on earth, good luck finding a worm. You can't do it. What does JJ do? He takes the machete, looks at his foot, cuts a slug of callus off of his heel because he's got this thick rhino skin, puts that on the hook, catches a six-inch fish, chops it in half, puts it on a bigger hook. And in 15 minutes, he's got a four-foot giant catfish that could feed a family of 16 and he's happy. I'm sitting there" }, { "from": "human", "value": "and I'm trying to stick a beetle on a fishing hook. Do you have just a line and a hook or" }, { "from": "gpt", "value": "is there a rod too? Just a line and a hook and then you just chop a rod and tie it to the... You just chop a little sapling. So are you still able to start a fire or no? I like for the food that I bring to not be fire dependent. And so if I have some nuts, I can shove in enough calories to get me through the night and leave a fishing line out and there'll be something there in the morning. But yes, I can start a fire, but a lot of times what I'll do is I'll bring a flask and not with alcohol, but with diesel. And so you have a tuna can and you put the diesel in the... This is what the local guys do. Everything I do, I'm sure there's gonna be someone listening to this and they're like, how could you do that? And it's like, this is what we do down there. Sorry. It's a tuna can, you pour a little bit of diesel in it, it burns slow, you light it and you put your sticks, you make your pyramid over that and eventually that will burn through the moisture and finally you'll get a very reluctant little fire enough to make yourself like a cup of tea" }, { "from": "human", "value": "or to pour that into the noodles, something, something. Or you just eat a fish raw. How important is it to stay dry? Is it basically impossible? It's impossible to stay dry. You're wet all the time. You're wet all the time. What does that mean? That means infections are more efficient?" }, { "from": "gpt", "value": "Yeah. So yeah, I don't know if you saw the picture in my book where I have the..." }, { "from": "human", "value": "The yellow spots? Yeah. So yeah, there's a picture with your entire face consumed with yellow spots is basically, I guess that's MRSA. Oh boy. So how did that happen? What was the infection like? Yeah. And how crazy are you for letting that infection stay in you for a prolonged period" }, { "from": "gpt", "value": "of time without treating it? Or you had no choice. So how did that happen? No, I did have a choice. I was 19 years old and I was taking care of a giant anteater that was orphaned and this is like my dream animal and she was mine and my job was to teach her the jungle. And so when I started noticing that I had an infection and that I had, I think I had dengue at the time too, I went back to town, probably picked up MRSA in the hospital where I got tested. Came back into the jungle and then got progressively sicker and sicker and weaker and weaker as I was two weeks, three weeks in the jungle. And it got to the point where my vision went black and white and I passed out one day. And I don't know why, but at the time I had shaved that day. And when I woke up the next morning, I couldn't open my eyes because the pus had come out of my eyes and out of the pores in my face, all those little micro cuts and the pillow was stuck to my face. And I was stuck up river with no help at 19 years old. And also when you see that picture, you can imagine that I assumed that my life was over because I didn't know what it was. And I also didn't assume that, or at the very least I figured I'd be disfigured the rest of my life. I didn't think there was any getting better from that. And so I remember sitting by the side of the river praying that a boat would come by, but it was the rainy season and there aren't going to be any boats because the river is psychotic. And so it was a long time before I got back to town and I didn't want to leave the anteater, but it became like, I was like, I realized I was dying. And then I finally got a boat with some loggers, a death boat, just loaded with these guys were, had gone into the jungle and shot everything they could and taken all the babies and they were gonna go sell them. So it was like baby monkeys and toucans and birds and cages and pieces of crocodiles and anacondas skins and jaguar skins rolled up and it was just like, I was just laying there with all these dead animals in the boat with all the flies on my my face and got back to the hotel, called my mother, said, please book me a flight out like today, like today. And then I sat on the plane and somebody sat next to me on the plane and I had a hood on. And I do remember that in the haze, at this point I was having trouble staying conscious, but I do remember that she like looked over, like trying to see what was sitting next to her. And then she got up and never came back. And when I got to immigration in New York, you know, the cop, he like takes my passport and he goes, yeah. He goes, so what were you doing in Peru? He's looking down and he goes, yeah. And he like holds up the passport, he looks at the passport, looked at my face. He goes, bo, buddy. And I said, no that's why. I was like, I'm trying to get home, to go to the hospital. He goes, he stamped it. He goes, go, go, go, go. God bless. God bless. He's like, oh, shit. And then they put me in the room in the hospital with like the hazmat suits and they didn't know what it was. And I spent like five days on IV antibiotics with like four different things running through my veins. And the doctors were like, don't let it go that close." } ]

[ { "from": "human", "value": "They're like, you went real close on that one. Yeah, that's what that picture, I mean, people should check out the book just to see the picture, because I imagine you just laying there, unable to see, have a fever probably. So you're like half hallucinating. Yep. And there's no boat, there's no way out. There's no help coming. Plus there is this creature who you've become a parent of, that you love. Boy, that's a dark place to be as a 19 year old. I mean, most people will never be in a place like that. Like where did you find strength in that place?" }, { "from": "gpt", "value": "I don't know, I just remember writing like a goodbye letter to my parents because they said if I die out here, it was really dark. Like it was terrifying. It really felt like it was the end. And I was writing, if you find me out here, I'm sorry, and all that type of stuff. And it was, I don't know about strength. There's no strength, it was just like move forward. And at some point it was like, if you'll take me down River, take me down River, and you know." }, { "from": "human", "value": "And you just got lucky with the loggers, with the death boat that they found you." }, { "from": "gpt", "value": "Well, how did the infections start by the way? What? I don't know, I really don't know. I mean, we always have some sort of little shit, but the thing is now JJ taught me that there's like three different trees that can cure infections. I didn't know this at the time. I didn't know the cheat codes. Now there's, if you have a small infection, you can use Sangre di Drago and it'll cure it right away. Like we'll see, you have a bot fly and it gets a little pussy. There's a fly living in your skin. You put that on there, not only will it kill the fly, but it'll heal the infection. Now, if you have a worse infection, you can go to Ojai, which is Ficus insipida, and you can use that. And that will completely heal, that will murder, it's like crocodile blood, it will murder infection. So like, forget neosporin, that's a joke. These are heavy chemical compounds running through these trees, and they know all about them. And so, whatever it is. So now at this point, that's no longer an issue, like because we know how to handle it," }, { "from": "human", "value": "at that time, if JJ had been there, I would've been fine. Well, learn the hard way. So these are open wounds, and then there's creatures that start living in them." }, { "from": "gpt", "value": "It's basically, what is it? Well, that's separate, that's bot flies, yeah. There's a creature that unfortunately, very, very, very unfortunately, likes to make its home inside the flesh of mammals. And so, the flies attach their eggs to mosquitoes. The mosquitoes go and seek out warm-blooded animals. The eggs, microscopic eggs fall into your skin and then begin to grow, and sooner or later, you feel a twitch. And it's a worm living inside of you that's vertical down in you, and it's eating you. And at first, it's not a problem, but when they get to about as thick as that pen, it starts to hurt, because you've got a hole in you, and they have a little breathing tube that comes up, and they breathe, and they go back in, and then they eat, and they come back up to breathe, and you have a friend living in you. They've had one of those? I've had lots of those. J.F. Alderman It's tough to take them out. They have hooks." }, { "from": "human", "value": "R.B. You gotta love the jump-up. So how do you take it out?" }, { "from": "gpt", "value": "J.F. You don't have to come. R.B. You gotta love the jump-up. So how do you take it out? J.F. You don't have to come. R.B. A hundred percent. J.F. You gotta put an irritant so like a lot of times what we'll do down there is someone will take a massive drag of a cigarette and then they'll spit the...they'll power like exhale and get some of the tar, which also shows you how much tar you get out of a cigarette. And then with a knife, you put that right over the hole and then you slap some Vaseline or something on top of that so they can't breathe. And eventually over the course of a few hours, they'll come up enough looking for air. And then you gotta grab them with a tweezer and try not to rip them, because then you're going to get an infection. And you gotta squeeze from the... It's a whole ceremony. When people have bot flies, we're all like, oh, it's bot fly time. Let's Let's go and then like JJ will squeeze. He's got like pliers for thumbs. He can like take a piece of your neck and you think he's going to break your skin. He'll just squeeze until this thing comes out." }, { "from": "human", "value": "And you don't wanna, I guess there's an open wound right there." }, { "from": "gpt", "value": "Yeah, you don't wanna bathe for a day or two after until that closes, because otherwise you're gonna have like water sloshing around in like a little pocket of yours. This is kind of gross. And that water might have other organisms there. Water in your skin tends to, yeah. I mean, the jungle water is clean. We drink it like I drink the water fresh out of the stream. Oh, that's interesting. Well, it's just a giant filtration system. All those roots," }, { "from": "human", "value": "the whole jungle is constantly purifying everything. People might be thinking about that with the jungle. There's insects probably all over you all the time." }, { "from": "gpt", "value": "It's not as bad as you think. Like I've been to Finland, Lapland in the summer and the mosquitoes are horrendous. Like devastating. The Amazon in our area, if you're sitting in a hammock reading a book out, our research stations don't have walls or anything. You're good for about one mosquito every half hour, which really is not a lot. I mean, it's worse in New Jersey." } ]

[ { "from": "human", "value": "Like it's really not that bad. Tell me a little more about the little baby anteater, Lulu. Lulu. That you've, who you've rescued" }, { "from": "gpt", "value": "and had to sadly leave behind. Yeah, I just was always fascinated with giant anteaters or this German shepherd sized thing with Wolverine claws and these giant Popeye forearms and they excavate ant and termite mounds and they had this long tongue and their babies right on their back for the first six months of their lives and so they actually have this incredibly intimate relationship with their young. And it just so happened that this animal that I was wildly fascinated with, there was an orphan on the river and JJ was like, you love these things. And I was like, yeah. And so he went and he's like, hey, my friend, you should, he got me the baby. And we were like, we're gonna rewild her. And so I spent like weeks and weeks and weeks just like with this thing on my back, crawling through the jungle, teaching her to find ants, giving her milk, falling asleep with her on my chest. And their tongue is like 11 inches long at that age. And so when she wanted me to wake up, she'd fired up my nose and it would come out my mouth. And then if I tried to get her off me quick, she'd stick the claws in and all my clothes, I have old, now they're like museum pieces" }, { "from": "human", "value": "with rips in them from Lulu's claws. Was able to also communicate emotion" }, { "from": "gpt", "value": "and feeling and all that. She needed it, she needed it. So if this animal didn't have the physical touch, if we didn't, if I didn't hold her all day long, she'd throw tantrums, she'd go shred something, she'd go pull down the curtain, she'd go ruin the woods, she'd start literally having a traumatic response to not having intimacy, which was shocking. Because again, on the scale of a cockroach to an elephant, you go, I didn't know that giant anteaters had such intense emotions, but she did. And also taking care of her forced me to explore the jungle from the perspective of an animal. So I got to be an animal. And so there's only a few times in my life where I've gotten to do that. One was with her, another time was living with a herd of elephants where I had to walk with them through the forest and see how they interacted completely natural. And it's different. It's very different. And you realize just like a person's public persona when they're out on the street in Manhattan is gonna be very different than when you're on the couch with them on a Tuesday night. And with wild animals, it's very much like that. If we see, you see a bobcat on a trail and it's gonna look at you and glare at you and then go off and it's like, yeah, but what's it like when it's in the den" }, { "from": "human", "value": "is playing with its cubs. Yeah. So that when it's looking at you, that's like the Instagram post it's making. The actual- It's doing duck face. Exactly. Yeah. So you've, besides Amazon, you spend a lot of time in India. Can you tell me what you learned hanging out with a herd of elephants? Yeah. What should, what do people not understand about elephants? That's beautiful to you. That's interesting to you." }, { "from": "gpt", "value": "First of the actual- It's doing duck face. Exactly. Yeah. So elephants should have government representation as like a subset of society, like actually. They- They have intelligence? They are so intelligent. And when you look at an elephant, so there's this question that keeps coming up of, are we smart enough to know how smart animals are? Can we interpret the intelligence that we're seeing? And I've lived with a semi-wild herd of elephants in India for a while. And some of the things that I saw like changed how I view reality, to be honest with you. Cause you know, you watch a matriarch of an elephant herd walk up to someone that none of us knew was pregnant and her trunk goes to her stomach, and then she calls all the other ones over and they're interested in this little human that they know that there's something in there, and they're all conversing about it, and you go, whoa, or that every morning we'd wake up and the elephants didn't want the stream water, they didn't want the lake water, they didn't want puddles. They wanted the water from our well. we had like a stone well, like a traditional and every morning it's like run out of bed because all the elephants were gonna come and they were gonna rip the bucket off and destroy everything. But they wanted that nice cold, clean water. And so it was like caring for elephants that were wild. That was sometimes getting shot at by farmers because if they went to try and rob some bananas, so these are sort of like delinquent elephants that were half wild and the forest department was thinking about getting rid of them which whatever that meant. And I made really good friends with this one elephant and his name was Dharma. And Dharma, this stuff doesn't, it's hard to write the book I'm writing right now because none of it sounds real or that. He grew up around people because he was a tuskless male. So he couldn't hang out with the females because he was a grown up male. And he couldn't hang out with the males, the bulls because he couldn't defend himself when they rough housed and everything. So Dharma would be like wandering around the forest not knowing who to hang out with. And so there was one night, there was a tiger calling and we just heard, you could hear it echoing over the hills. And what does Dharma do? 2 a.m. we hear Dharma show up. And he's same thing, he starts throwing a tantrum, he starts pulling shit over, he takes a chair, throws it. We had bananas in the truck. Dharma walks up to the truck, it's like a Jeep. He walks up to the Jeep, smells it, he looks at me, and he's like, are you gonna get out of bed? I'm like, no, I'm not gonna get out of bed. I was like, Dharma, you're a grown-ass elephant. The tiger does his thing again and he's like, I need bananas to feel better. Pushes the truck up on two wheels, looks at me. Is this how you want it to be? So I'm up, I'm up and I go and I'm like, please, please, please, please, Dharma, I'm rubbing his face. And he's like, he puts it down, he's like, all right, well then, then hit me. I didn't do it, so he lifts it up again. And so in the end, there was no way for me to outsmart the elephant, he wins. There was nothing I could do. And so a lot of my job was taking him out into the forest and spending a little bit of time with him. I have this beautiful, one time I set up the tripod and I went and I was just, I was just journaling. And he would come and he would just play with my hair and he'd be like, hey, what's up? He just, he wanted someone to interact with on an emotional level. And when you think about elephants in terms of the fact that people go, oh, they use medication to induce labor, it's like, yeah, that's not that surprising. They hold the bones of the dead. It's like, yeah, they have the best smell of pretty much any animal. That's also not surprising. They probably know exactly who that was, that bone. But they can navigate to water holes and communicate in ways that we cannot really figure out. And so when you hear about people measuring elephant intelligence, you'll hear about scientists being like, oh, well, we gave it a, you know, a bucket with a hole in it and then it had like a key and there was a rope. And you're like, bro, this is all human stuff. Can you go walking with them for three weeks in the wild and watch how they deal with the problems that they encounter in the forest? And so elephants have become, especially recently with the work that I've been doing in Africa with Betpaw, I've just become so fascinated with elephants and, you know, elephants, the African elephant population right now is down at 2% of what it was a few hundred years ago, we're really putting them on the brink of, you know, there's some elephants that are being born tuskless because we've, poaching has taken down the great tuskers to the point where now it's actually beneficial for some elephants and not have tusks cause they won't have humans, but that's like, we've created deformed elephants. And so like now I'm kind of very concerned" } ]

[ { "from": "human", "value": "with issues of elephants and tusks" }, { "from": "gpt", "value": "are fundamental to the interaction between the elephants. Absolutely, I mean, with males compete with each other, but also elephants use their tusks. You know, like they'll, they'll break a branch and they'll be like, this is a good branch. I'm gonna eat the hell out of this, and they'll like hang it on their tusk, and they'll like grab a bunch of other stuff, they'll like hold it, ripping a tree up out of the ground. I just watched it, two weeks ago as watching an elephant, he got down on his knees and stuck his tusk into the ground and like leveraged up he like arcametes to this root out of the ground and then was like, that's a sweet root. And then when he left, I went and I tasted the root and it was like sweet jinder and I was like, I have no idea what this is" }, { "from": "human", "value": "but he knew it was good. Do they use tusks for sexual selection," }, { "from": "gpt", "value": "like to impress the ladies or no? It's certainly involved in who has mating rights. Oh, who wins in the competition. Who wins. I mean, if you got the big tusks, and there are elephants out there, like the mammoth big tuskers, that have tusks down to the ground, like huge. And when you see them, it's like seeing something unique on earth, unique in history, because we're at a point where we might lose those. There are only a few of them left." }, { "from": "human", "value": "And then they're so prized by hunters. Yeah, it's interesting, but I forget what the actual conclusion on that is, because there's some studies of the use of, the value of beauty in evolution, like birds and peacocks and so on, there's no actual value to it, but it plays a role in sexual selection, meaning value, like, it's much easier to understand competition, like a tusk helps you defeat the competitor. But I bet you there's a component to the tusk where the ladies go, God damn, it's a nightmare. Like there's a visual, beautiful component, maybe not, I don't know." }, { "from": "gpt", "value": "But what if beauty though, as we're defining it though, is symmetry and the absence of yellow spots on your face and healthy looking hair. And so like, I think to us, beauty is sexually appealing traits" }, { "from": "human", "value": "that look good to mate with, and so. So that 19 year old with Marissa, everybody in the world will swipe left on that. The least desirable object in the universe. Okay, speaking of elephant intelligence, it's something I think and work quite a bit on with artificial intelligence, is what the philosophical question that comes up is, what is intelligence? What is intelligent? Humans, homo sapiens are often thought to be highly intelligent. That's the reason they stand out. So in your understanding of different species like the elephant, what stands out to about humans? Or are they just another animal" }, { "from": "gpt", "value": "with different kinds of intelligence? Well, we're certainly unique because we have altered the entire planet. Yeah. You know, the term the Anthropocene. I mean, it's like we've literally created a geological layer of us. Whereas other animals don't, and going back to elephants, it's like, they also engineer their environment. If you're in a forest, like if you drop me in a forest on earth, I could tell you in two seconds if there's elephants there because there's twisted branches and excavated earth and they're constantly gardening. But I mean, look at us. I mean, we're clearly unique in nature, which makes me not understand the anti-human sentiment that so much of environmentalism has about we're bad, we're damned, we ruin everything. And it's like, I've seen the worst. I've seen the burning Amazon and I'm still like, I love being able to share ideas with you and travel to places and FaceTime my family when I'm not around them. And it's like, I celebrate a lot of what makes us human. And it's almost like reality is this crazy video game. And it's like, if we could just figure out the right keys, we can pretty much do anything we can think of. And it's like, I mean, poetry, art. I mean, I'm the biggest animal lover in the world," }, { "from": "human", "value": "but we are different, we really are. Yeah, the ability to puzzle solve, create tools. We think it's the coolest invention humans have come up with. Is it fire? What's the most impactful?" }, { "from": "gpt", "value": "Fire is fire. I feel like fire is kind of a gimme. I feel like the- They didn't really invent it. They probably-" }, { "from": "human", "value": "Like the wheel." }, { "from": "gpt", "value": "Invented, they probably- Flying. I mean, flying. I mean, think if you could go back in time to someone that never flew. You know, a sultan, an Egyptian king, George Washington. You know, and be like, you can fly. I mean, just on my way here. And I fly way too much, but I was looking out the window at the clouds and going, this is unbelievably spectacular. It's just stunning. You know, as a kid, you look at a cloudy day and you go, the world is like this today. And then you get in a plane and you fly above the clouds and it's sunny up there. And you go, oh, it just, it changes your perspective. It's like when people go to the moon and they come back and they tell you the pale blue dot. You know, just, I say flying. I think the ability to fly. I mean, the fact that I could get on a plane" } ]

[ { "from": "human", "value": "and be in India in, you know, 22 hours is shocking. In terms of its usefulness, I would argue that's not in the top five, but in terms of its ability to inspire. Somebody I forgot who told me this idea that there's something about the atmosphere, earth's atmosphere that allows you to look up and see the stars. Like if we didn't have that, human civilization would not have happened. Meaning like being able to look up and see something out there would fill our, like something that allows you to look up versus just look down. To like first looking at your local environment. Be able to like wander and see, holy shit, there's a big world out there. I don't know anything. If you're able to look up and see that, that kind of humility combined with the ability to dream about exploring. Maybe it just inspires the exploration. It's kind of an interesting thought given how inspiring, for example, the extra upgraded super cool version of flying, which is flying to other planets. I mean, there's going to be, hopefully it's possible. This century, a child born, no, not this century, maybe this century, a child born on another planet that looks up, that looks back at earth. And it has to be educated by his, her parents that like there's another place, there's another place" }, { "from": "gpt", "value": "where life is way easier. Oh God, it's so easyattering. There's water everywhere." }, { "from": "human", "value": "Exactly. Yeah, exactly. People complain about earth, man," }, { "from": "gpt", "value": "the earth is really good. It's really, really good here. Water everywhere, anything. I wouldn't even leave, given like right now, like if somebody said, like, oh, you could like, you could go to the moon. I'd be like, no, I'm good. If I died in space, I'd be so pissed." }, { "from": "human", "value": "I love it here. Yeah, but you're still," }, { "from": "gpt", "value": "there's a longing to explore for you. There's a longing to explore, but I really think I'm such a, like my longing to explore is like rivers, streams, oceans, jungles, like to me, yeah. I would watch the hell out of the live stream of Elon touching down on Mars. Like I'd be like, this is incredible. It's amazing that I get to be around to see this." }, { "from": "human", "value": "I'm staying where I'll be right here. Yeah, but it's good that the human spirit pushes us. Oh, it's amazing. That was possible. And it does that for you. What, just the out there questions, what's the most dangerous animal in the Amazon, would you say? Mammal, let's go with mammal. Dangerous mammal. Like dangerous in terms of you walking around, doing the solo hike." }, { "from": "gpt", "value": "I'm gonna disappoint everybody with this, but it's humans, it's nothing. There's no, if I'm out in the Amazon, there's nothing that's going to attack me. You know, and in India, you might have an old leopard or a tiger that's missing a tooth that decides your prey, or you might have an angry elephant that's in must that just decides to flatten you. In the Amazon, you're not, there's real, jaguars won't even let you see them. Yeah. And there's really nothing else. One of my friends, a brilliant scientist friend of mine, Pat got attacked by a rabid ocelot once, but that's like a dieseled house cat just having a fit." }, { "from": "human", "value": "You know, it wasn't the worst thing in the world. Yeah, it's just the assholes of the animal kingdom. Okay, what, in terms of humans, you said that the tribes, some of them, uncontacted can be exceptionally dangerous. What's your experience with them? What should people learn? Because it's such a fascinating part of life here on earth that there's tribes that don't have much or any contact" }, { "from": "gpt", "value": "with the quote unquote civilized world. Yeah, most of the people that I meet don't actually really understand how isolated these people are or how weird it is that we're sitting here and that we have iPhones and airplanes and all this stuff. And these people are living naked in the forest at this moment. And so the thing though, you know, I also was recently somebody said, somebody said, oh, there's like paleolithic tribes. And it's like, no, no, no, just by default, they're modern tribes living now. They just happen to be living out in the jungle. And there's a huge debate about, you know, do we try and contact them and bring them in? And there's two camps of people on this who they say that it was the trauma of the rubber boom that sent them out that far into the forest and made them terrified of the outside world. And so that's also what made them so hyper violent. I mean, there's one of the guys we work with on our team, Victor was in, I think it was 2004. He's coming down river and he had a load of mahogany wood and he's piloting this boat and he sent two people, husband and wife ahead to go start cooking breakfast on the beach so they could put the little kitchenette thing down and pick, put the propane. He sent them ahead as the, he's going nice and slow with the barge coming down the river. They go ahead, reach the beach, they get out. He starts cutting some cane to start making a fire. Tribe comes out, no warning. They just start screaming, they start shooting arrows. The man instantly gets an arrow through the leg and it pins his leg so he can't run. He tells his wife, go save yourself. And she does, she jumps in the water. There's arrows falling around her too. And as she's floating down the river, she looks back and the last thing she sees is these guys getting to her husband and beginning to rip him apart. As Victor comes down the river, this is a guy we work with every day, he comes down the river and sees his friend disemboweled, opened up, dissected, his parts are all over the beach. The beach is red and they only found out what happened because they found her later on holding onto a stick in the river and they were like, what happened? And she was like, they just attacked. They don't want people on their land. On the sort of the underground WhatsApp chain of the Amazon, they, a few in August, like this was not internationally known some loggers went up and tried to steal a few trees from where the tribes were. And then everybody sent the pictures of what the loggers" } ]

[ { "from": "human", "value": "looked like after a few days because the tribes around, that they construct around who these outsiders are. Are they gods? Are they demons? Are they humans? Who are they? Who are we to them?" }, { "from": "gpt", "value": "Well, you got to go back to the rubber boom. The rubber barons went down there and at the start of the Industrial Revolution, the only way to get rubber was to mine it from the trees that were out in the forest. So the only way to do that, because you can't make a rubber plantation in the jungle. When it's in plantation form, when it's a monoculture, culture, it gets this leaf blight and it all dies. Henry Ford tried it, it didn't work. And so what they did was they sent these people down who just whipped, burned, enslaved, raped, and pillaged the people. It's one of the worst periods in human suffering that I've ever read about. One missionary said they were killing the locals the way you or I would kill a mosquito. They just went nuts. And so they sent them out and they would come back with rubber, and this would go to fuel the industrial revolution for hoses and gaskets and tires and all this stuff that suddenly we needed. And it was during that time that these gangs of foreigners would go into the jungle to enslave the natives, that these uncontacted tribes went back into the jungle and said, not us. And they have six foot bows and seven foot arrows with bamboo tips. They make the bamboo tips into razor blades. And so when those things fly, actually one of my rangers, one of the Jungle Keepers team, was present when the tribes had come out onto the river and he tried to help them because they're nomadic and they live out there. And so there's an element of like, brother, like, you know, they're trying to be like, you don't need to be like this. Like, we're friendly. So they sent a canoe across the river with bananas. And so he's up to his waist in the river and the tribes are right across the river and they shot. And he sees the arrow coming right at his head. And as he moved to the side, it hit him at the temple and sliced him back towards the ear, opening him to the skull. He's fine. But let me tell you something. When he goes and gets a crew cut, it's the most badass skull, the scar you've ever seen, man. And so he always keeps it real short on that side. But even if you try to help them, they're not necessarily friendly. Tough. That's a tough lesson. Yeah. I suppose they have a point. They have a point and protecting them is a default of, you know, now that we're protecting all this ecosystems and all these other indigenous communities, it's like we all sort of live with this knowledge that the Hermanos, the brothers are out there and that's the way they want to keep it. And so we just have to be respectful of like, you don't camp on certain beaches at certain times of the year because we know that they might be there. You really have to be careful about that. Have you yourself interacted with any? My interaction with them came on a solo where I pushed it a little bit too far. And I was planning to do a three week. This was like the big one. And I got dropped off by poachers up a river and I went past the point where they were like names there. I said, what tributary are we on? And they were like tributary. And I was like, okay. And I said, leave me here. And I remember the guy being like, are you committing suicide? And he didn't understand that. I was like, no, I have a backpack and I have food and I'm going to take videos and have a tripod. And I was like, we're cool here. And they looked at me like they were like goodbye. And I was like, all right. And I went up this river and again, you learn these things. It was only when I'd been alone for a week that you realize you're, what's that saying? They're like, oh, you're born alone and you die alone. It's like, no, you're not. You're born into a room full of people usually. At the very least your mother's there for everybody. And so you've been around people probably, if you're a normal person, every single day of your life. You've seen dozens, if not hundreds of people. And all of a sudden you realize what a social creature we are. Because on day six, it gets weird. For me, it got weird. I know there's people that can do it longer." }, { "from": "human", "value": "So what does that mean? Like longing for contact? Like lonely?" }, { "from": "gpt", "value": "Lonely. Longing for contact, the distortion of reality in the sense that like, you wake up and there's no one there and you start to, you're going up a river. So I kept looking back down river and almost thinking of my life as something, it was almost like I had already died and I had gone to somewhere else. And I was looking back on that life as like something that I had experienced. And then there came this panic of, what if it's gone? Or like, what if World War III broke out and I just don't know about it. My family in New York is vaporized. And something just, you just," }, { "from": "human", "value": "your ability to comprehend and interpret reality kind of requires other people. It's not just that you're lonely. You need that contact to actually just perceive the world, make sense of it, all of that. So you start basically hallucinating in a certain kind of way." }, { "from": "gpt", "value": "I started feeling very uncomfortable. It doesn't help also that like Santiago told me these stories where he's like, if you hear Cappuchins sounding, not quite monkeys, if you hear Cappuchin monkeys sounding not quite like Cappuchins, he goes, it's the tribe and they're coming to get you. And then the guy who was shot, Ignacio, they showed me videos where we saw them on the beach and they're communicating in monkey calls. They're using it as code so that we don't understand them, even though we don't speak their language, but they're using animal calls. And so, every night you go to sleep and then you go, did that tinnamoo sound off? And you're like, shit. And it's really hard to fall asleep. And then one night I messed up and I left a fish. I cleaned this fish. I ate this huge fish. I just ate it to my face. You're putting out marathon levels of energy every day. Goggins would love solos. He'd be like, this is awesome. Yeah. Do you eat fish raw? This one I actually cooked it, but the skeleton was laying there right outside my tent. Stupid. And in the middle of the night I wake up and I just could tell there was something there. And then you almost don't want to look. It's like when you're a kid at the basement door and you're like, is there a ghost? I unzip the tent and I open it up and there's 27 black caiman outside of my tent all looking at me like this. And some of their heads are this big and they're like, there's fish there. Can we have it? And I'm like, holy shit. And I had to scooch the tent back and move back and let them have their fish. And there was a host of crocodiles outside of my tent. But no, so then there was like 27 maybe. There was a lot. Big ones, small ones, medium sized one, every type. They were all there and their eyes glow in the night. You shine a light at animals and they have a tapered emlucidum. And so their eye shine comes back at you. If you shine a headlamp at a jaguar or a frog, almost every animal has a tapered emlucidum." } ]

[ { "from": "human", "value": "Wait, how many? See the croc, there's a whole lot of them. I thought, can we go back to the farther conversation where you said the jungle is not dangerous. The humans are the most dangerous. Well, did they eat me? No. Why didn't they eat you? They wanted the fish. Is there some way of you interacting with them that shows that you're not social?" }, { "from": "gpt", "value": "Well, did they? I don't believe so. I'm sure there's someone out there that thinks they can" }, { "from": "human", "value": "talk to crocs, but. Because there's a story of you grabbing a croc by the tail. What did you learn from that? Learned to not always listen to JJ. So JJ was testing you to. Yeah, to see how stupid I was." }, { "from": "gpt", "value": "How do you hold the croc out exactly? You have to get them by the head, like an anaconda like this. And so." }, { "from": "human", "value": "So you're one of the world experts of grabbing creatures by the head." }, { "from": "gpt", "value": "I wouldn't say world expert, but I've done a lot of it. I also have, you see how there's kind of a ball there? That's where a crocodile tooth went in that side and came out that side of my mouth. That was a really good chomp. And the watch I was wearing at the time saved me because that, like that. Real fast, just chomp. Just whack. Like somebody took a sledgehammer, you put your hand on the table and I just whack. Really hurt. Shouldn't have been doing that. How did that come up? Because I caught a croc that was too big. So usually when we catch little caiman in the streams and we measure them to monitor the populations, you get it by the neck. And then I tucked the tail under my arm and I hold it. And you're talking about a little, you know, four foot croc, nothing. And I, this one, I dove into a swamp and I caught like a six foot spectacle caiman and her head is big and I had her by the neck and I realized I couldn't get her tail under my arm because her tail was all the way back there. And she started thrashing and it was like probably croc number 375 that I'd caught. And I just got a little cocky and I said, ah, she's, you know, I just, I just like grabbed her by a leg. I was like, I got this. And she just came back and tagged me and I went, okay, gonna go back to being safe." }, { "from": "human", "value": "Just to linger on it. What, is it one of the, one of the bigger predators in the Amazon? What, and it's, is it going to, are they going extinct? Black caimans?" }, { "from": "gpt", "value": "Black caiman were, I believe they were critically endangered for a while because for a while the fashion industry loved their skin. It's soft and it's black. They're bouncing back a little bit now, you know, like most animals, if you leave them alone, they'll be fine. I mean, crocs have been through, you know, how many millions and millions of years on earth before us. I mean, that's even the joke with, with the joke, but that's the grim reality of tiger conservation. It's there was a hundred thousand tigers in 1900. Now there's 4,000 tigers left on earth. It's not rocket science. All you have to do is not bulldoze their forest and allow there to be some deer and tigers will be fine. That's it. It's so simple. And that's like sometimes where I feel like I have the dumbest job in the world. I'm like, guys, please stop killing the things that keep us alive. The Amazon regulates our global climate," }, { "from": "human", "value": "produces medicine, is home to indigenous people. It's beautiful. Rainforests only cover 3% of the planet's landmass. Like it's not that much to ask. If you leave their home untouched, they'll figure out how to have sex and multiply," }, { "from": "gpt", "value": "except for pandas, apparently. Because pandas, you have to convince. Yeah. Humpbacks, humpbacks, they went down to, they went from 130,000 down to, I think about 8,000 at whaling times. And then when we banned whaling since that time, where I think we're back up to over a hundred thousand humpback whales, they've bounced back. It's a success story. We're not gonna lose them. Okay. So, you're on a solo with, with the crocs looking at you. So this is why you're good at this, you know. I would have lost, we would have been." }, { "from": "human", "value": "No, that's pretty epic. With the fish, that was your mistake? Okay. With the fish?" }, { "from": "gpt", "value": "When you come, we're gonna, I'm gonna show you." }, { "from": "human", "value": "You told me you're coming. You told me you're coming. 100% coming. If there's any place, I mean, sort of a grim joke, but if there's any way to die, that's a good one, if I'm being honest. It's a cool one. It's a pretty cool one. It'll become part of the- Part of, yeah. I mean, I'm not even joking. There's a oneness to the whole thing. All the" }, { "from": "gpt", "value": "stories, just reading your work, looking at your work, it's a pretty cool one. It'll become part of the-" }, { "from": "human", "value": "Part of, yeah. It seems like you're part of this machine that is nature, this incredible machine. We all die and we're all part of this big thing that humans do have the capability to also construct narratives and stories and myths and tell them to each other and share them with each other and have more sophisticated ways, therefore, to communicate love to each other. But animals do as well. They communicate love, maybe more simply, maybe more honestly." }, { "from": "gpt", "value": "Anyway, so you're with the crocs and the fish. Yeah, so I messed up. I left the fish out. Crocs showed up at some intent. But in the end, it was fine. I backed off. They had their way with the fish and then they all started biting each other. It was fun to watch." } ]

[ { "from": "human", "value": "Is that a general rule you want to not leave?" }, { "from": "gpt", "value": "Yeah, just like if you're camping in the Northeast, you don't leave. You do a bear bag or a bear canister. You don't want to invite the wild animals. I really did mess up. I was just like, whatever. I do this every now and then. I get a little too cavalier and the ocean has almost taken me down for that a few times. But yeah, so the crocs, and then you keep going for a few days. My plan was to get to a point where I reached the end of the tributary. Again, for me, this is like a pilgrimage. This is like me going into the heart of the very center and soul and essence of everything that I am fascinated with, as close to God as you can get, because you're leaving every type of security, every human relationship. You're also pushing all your chips in. Every step I took further up river, it got weirder and weirder and more intense. Every day and every moment it changed. I brought pictures at the time. There's no way to keep a phone charged. I didn't have a power bank or anything. I brought pictures from home. I brought a National Geographic magazine, something just to... And there came a day right when I was getting to the end, to the point where the river was so shallow that it was just a trickle and I was walking on the rocks and the Andes Mountains were in front of me and I was reaching the place and the music was swelling and then all of a sudden I saw smoke around the next bend. My spine is reacting right now as I talk about it because I knew what I was gonna see because I knew that it was impossible for loggers to be out there. There's no motor that could take you. The boat would have run aground miles ago and so I went and this is the other idiot thing. It's like just turn around. Just do it. I'm that kid though when you see like a wet paint sign, like I walk by and I touch the wall and so I went around the bend and I see a few naked people on the beach and they see me and we're like a good distance apart. It's there on the other side of the river but arrow in hand, bow in hand, the intention of pose. They're looking at me. They're clearly conversing and that moment lasted for a long moment where I said this is the part of the story where they are going to rip me apart, dissect me to see what I eat. I mean every other story in the region that we've heard, that's the ending of it. If you're alone with these people, it's not gonna go well and I have nothing to defend myself with and I just, I turned and I ran for like three hours and I got in the river and I swam for a while and all my food got wet. I mean everything, I just, you know, all systems go. Just ran for dear life and my get-out plan, the thing after I crossed the mountains and came down into the next tributary was I had a pack raft. It's a tiny little inflatable raft good enough to handle rapids and I inflated the pack raft once the river was like six inches deep. I inflated the pack raft and I went for the rest of that day into the night. I went into the point that my headlamp died and I was just floating in a raft down the Amazon and hitting into things and I was I was like, okay, I'm gonna pop this raft. So I got out of the raft, set up my tent, and I was like, I need some sleep. I was freaking out. I hadn't had food in now, you know, hours and hours and hours. As soon as I fell asleep, my asshole brain comes up with the dream of that I hear voices, they're right outside the tent. I just, you know, sleeping was worse than being awake. So I woke up, got back in the tent, and then at one point it was really cool, because one of those, one of the same black caiman that had come for the fish as I'm going down river, he came right up next to me and the two of us were going, and he was just like motoring down river, this giant like 16 foot crocodile. He just like came up to me and like looked at me as I was going, and it was funny because I wasn't scared of him. I was scared of them." }, { "from": "human", "value": "All sorts of stuff." }, { "from": "gpt", "value": "Just ran for daylight. And yeah, it took me like a week to get back to town. And again, the things you learn in these moments, the, you know, the appreciation for your parents, the, what a hug feels like, you know, when you are faced with pretty much certainty that you're not gonna get those things again, whether it's from MRSA or Uncontacted Tribes or, you know. I find that it brings you this new joy for life where you- Just being that close to death." }, { "from": "human", "value": "Yeah, you go, my God, this is all a miracle. It's sad because they're human just like you. Actually, how different are they? Like if you were forced to interact for a week together where they can't, they're not allowed to kill you- Not allowed to kill me. What would they, are they," }, { "from": "gpt", "value": "how fundamentally different are they, do you think? I don't think they're that different. I think they're like any other Amazonian natives. They're tall. They seem to have tall genetics and there's places, you know, again, there's what is known and then there's what we know down there. Like, there's one community where, I don't know whether it was like a bad rainstorm or something, but some kid from the uncontacted tribes did end up in a village. And so he learned Spanish or he learned whatever dialect they speak in that village. And so he's told us a little bit about what life was like with them, but like they're just people, they're just people. They have their own culture. They know about medicines that we don't know about. They definitely have hunting practices that, that we don't understand. They can hit a Spider Monkey out of 160 foot tree with a bamboo arrow. We can't do that. I mean, they are incredible hunters and also like living naked in the jungle with the bot flies and the mosquitoes. They're different, I think. I don't know how they do it. Like sometimes at night, and again, we don't have night vision, whereas almost every other animal does. And sometimes we'll be sitting on our, at the research station at night and we'll be just drinking and looking out and at the, we'll scare each other. We'll go, you realize if they were out there right now, they could be looking at us and it's like, The truth is that when it's dark out there, they can't see. It's not easy to start a fire with matches and a lighter and gasoline, they do it with friction. They have some beads on survival that we could really learn from. Not to mention that then you have people that believe that they are actually the guardians of the extinct giant ground sloth. And what they're doing is living out there because they're protecting a secret population of previously extinct megafauna. But there's all kinds of, I mean, this, it's like, you go into the crypto world so quick." } ]

[ { "from": "human", "value": "I've heard so many people be like. But then again, you have to be humble at how little we know about that world, about the world of life. Like you said, there's so much of life in the Amazon" }, { "from": "gpt", "value": "that we don't, creatures with no names. Tons of them. We could go out on a night walk right now and I could show you something that, you know, I've done it. You pick up a bug and you go, that doesn't look right. That's not right. He's got three heads. And then you send it to the greatest expert on that genus of insects and they go, look, I got no idea. And you talking to a world expert and it's like, that's it. And then 50% of the life is up in the canopy. And so like we started climbing the trees, like rock climbing, like what Alex Honnold does, like, well I'll climb up 50 feet and then I'll put a safety. Like I'm basically trad climbing and then I'll climb up another 50 feet and I'll have JJ belaying me from below. And then he'll be like, ooh, look, a snake. I'm like, JJ, pay attention. Like, did you get up there and the branches are as thick as this table. So you can like walk around freely. It's like total like avatar when they're in the floating islands. Like you can go run around if that's what you want to do. Bremeliads, orchids, cactuses, because up there it's against the sun. So it's a different environment up there. Oh, wow. Yeah. Interesting. And then you start seeing lizards and snakes and birds and things that aren't down on the ground. Yeah. Interesting. And so how many scientists have actually gotten to really spend time up there and really inventory the life. And that's why when you hear about, you like it, it's like a taxonomical discussion of how many species there are on earth. They're like between, you know, 10 and 30 million. And it's like, well, that's a big," }, { "from": "human", "value": "it's a big swing. What about stuff on the ground? So you mentioned some insects. What about bullet ants? So it's supposed to be the most painful bite in the world. You've been bitten by one?" }, { "from": "gpt", "value": "Seven or eight times, yeah. What does it feel like? Okay. So the first time that we ever did bullet ants, JJ said, you know, okay, this is what we're going to do. He goes, you know, it's bullet ant roulette. We've gotta get a bullet ant and you ever get like chopsticks, you pick up this bullet ant and they're big. They're big and they're tough. Then he goes, we're gonna put our four arms together and we're gonna drop the bullet ant and clamp our four arms together and just rub and whoever, whoever it takes, it takes. Of course, JJ did not get stung and I did. and it hurts every bit as much as they say it hurts. It really let me have it. And then I was like hitting my arm against the table to try and like kill it or get it off, but it was holding on and just like really injecting the venom. And yeah, really letting you have it. And then it travels up and it goes into your lymph nodes and into here and you get a headache. And I think the brilliant thing about the venom of a bullet ant is that it makes you feel likes like this this feeling of alarm it makes you feel like something's wrong you don't just go out this hurts like it's like a nuestring where you oh this really hurts on my hand it's like no no no no your whole nervous system is freaking out and you start sweating and then you get cold and then you're tired and then you get a little blurry vision and it's like that's actually that bad i mean now after six or seven i get bitten and i'm like Kinda okay so it's a full-body full-mind experience it's a full-body full-mind experience but then there's places in the amazon Where they you know stick their hand in a glove with like 70 of them, right? I think Steve Oh did that which I would I just don't understand how you could do that without going into complete anaphylactic shock and dying because one" }, { "from": "human", "value": "Really sucks. Yeah, so it's a body. Well, just like just like we said with animals and with humans. There's a different kind" }, { "from": "gpt", "value": "The first of his species" }, { "from": "human", "value": "Um On the point of on contacted tribes, it's interesting to think about what kind of Civilizations have there been yeah, this is something that you've talked about a little bit. Graham Hancock has written about Ancient civilizations sort of challenging the Conventionals of the mainstream thinking about the the civilizations that have been there in the Amazon. Can you steal man? And criticize the idea So the pro and the con of the idea that there have been ancient advanced ancient civilizations in the Amazon Like how much do we know what what are all the possibilities?" }, { "from": "gpt", "value": "Of what's in the amazon in this past so like when oriana went down On the Amazon the reports were that there was great civilizations in the Amazon And then you know a few hundred years later when people got to actually check up on this stuff It was all gone And so was that because of disease that we wiped out all these civilizations and these communities of people potentially probably Was he just wrong? Probably not this is a guy that navigated by the stars back to spain after building his own boat like yeah Or did he you know, or was he trying to just I don't know. I don't know but There clearly is a long history of complex civilizations in the amazon 100 There's no one that can deny that the thing that I reacted to was that i've Heard videos i've seen moments in podcasts where The narrative becomes not there's more ancient civilization information in the amazon than we previously thought true statement We're discovering with lidar and this is what graham hancock is talking about that we're discovering constantly that there's There was more civilizations than we thought in various places The place where I take offense Is where they start to say that the amazon there's actually articles that that are titled this that the amazon is a man-made garden" } ]

[ { "from": "human", "value": "Which is not true so the actual Which I think is a really different idea that the uh The entire ecosystem everything we've been talking about all the species Uh all the forestry and the different Just life life one of the most diverse Ecosystems on earth is initially created by humans. It's ridiculous. Well, it's not first of all It's unlikely but it's not ridiculous. So we can't well, there's no ridiculous in um in science, but The complexity of life is very difficult to engineer as you the more you study about biological systems and so on It's very difficult to create the kind of things that nature is able to do That said I don't know if you've heard but the entire earth The world has gone through a pandemic recently And if and everybody said of course, it's natural origins viruses mutate all the time And nevertheless, it seems more and more likely than this particular case. It was of an artificial origin leaked from a lab So humans are able to create stuff. Sure these modern technological Genetic engineering things that need made golden retrievers. Come on. Come on. You can't be that nice and that good looking I used to be a wolf. Yeah, but so that bothers you because it it allows you to think that" }, { "from": "gpt", "value": "We don't need to preserve the Amazon. We can always engineer it. Yeah. Yeah, exactly Then just this is just to me. That's a slippery slope. Like I totally um, it's just it's so quick from i'm a fan of Expeditions to find ecological ruins and and to learn more about the ancient civilizations To which I don't think is what he's putting out is that Then sort of like news articles, which I think they're trying to bait you where they're going was the amazon man-made and it's like yeah You know because then you get a you're going to get a brazilian president to go see see what they said It's man-made so we might as well continue to engineer it and manage it and it's like there's such complex systems and interactions and such a such a giant web of life there that at least in my opinion is clearly one of the most Authentically natural things and again, are there things that we've engineered to the uncontacted tribes? Sometimes they they have banana plants that they've stolen and we can see it from the air that they have banana plants We've made banana plants that's engineered by us. We know for a fact So agricultural engineering agricultural engineering and stuff like that, but suggesting that the amazon basin You know, it's just it's just a weird way to think about it. I've just heard people dismiss the protection of the Amazon based on the fact they're like, oh, well, if people made it, and it's such a giant leap from zero to 100, is there slash and burn that the ancient civilizations did? Of course. Are there areas that were affected by people? Of course. I just get worried when we" }, { "from": "human", "value": "start talking about it was a man-made thing. Hear your engineering thing. Loud and clear on that. And I personally think that's completely separate from wondering about what the ancient civilization have been able to accomplish. Oh, sure. It's almost really sad because if all the humans on Earth die now, how long does it take before all signs of humans ever existing disappear? Ooh. For the most part, from an alien perspective. What timeline are we talking about? I mean, like 100,000 years? It could be less. It could be less. It could be like a few thousand because 100,000 is complete destruction. 100,000 is like nothing. But then it could be in just a few hundred years, it starts becoming the government of the alien civilization is going to have to pay quite a bit of money to do the research because they're going to find other life first. They're gonna find the dolphins in the fish and so on. They're going to find the trees. Maybe the trees are the interesting thing." }, { "from": "gpt", "value": "The buildings are not that interesting. 100,000's like the crumble, but there must be examples of cities that have been left unattended for a few decades and how quickly the plants push up through" }, { "from": "human", "value": "the street and everything starts to get broken down. If you really look you'll be like, oh, Oh, there's some interesting geometry here for the buildings and so on, but most of the computer stuff, all the stuff of the past hundred years, the airplanes, all of that, all the technologies, all of the paperwork, all the hard drives" }, { "from": "gpt", "value": "that store all the information, all the hard drives. I wanna actually know how long it takes like a 747 to like biodegrade. Like how like if you just leave it there, sitting on the runway, society stops. How long does it take for that thing to disappear?" }, { "from": "human", "value": "Like that's a weird- Certainly versus to a point where it's unidentifiable, it might be different, but- Sure. I mean, the point I'm trying to say here is as you've brilliantly put the Amazon churns. Yeah. Oh yeah. And the fact that I wonder throughout its history, what are the peaks of the awesomeness? Well, how many banana, how many agricultural Einstein's of bananas were there? They're creating different kinds of ideas," }, { "from": "gpt", "value": "different kinds of geometry, different kinds of tools. Well, yeah look what the Incas did. I mean, the Incas Machu Picchu, I mean, when Hiram Bingham found Machu Picchu it was covered in jungle, you could hardly see it. And I mean the stonework they did, much like what the Egyptians did with the pyramids, a lot of it we don't really understand how they did it. If you come to the jungle, you gotta go to Machu Picchu because it's not far from there. And I usually, I'm the person I won't... usually go see like the you know like I've never been to see the Taj Mahal after living in India for five years like I'm just not but when you look up and you see Machu Picchu you go either they were communicating with the gods there or these people were so smart that they knew that anybody they brought they were gonna impress they they've built something there that when you look up at that mountain you go whoa with those giant stones the beauty of it you know it's just it's just stunning to imagine that there was this culture of people that that could achieve this and so through the Amazon I mean that's sort of up in the Andes but there's all kinds of stuff in the Amazon there are places where they say there's pyramids beneath the canopy" } ]

[ { "from": "human", "value": "that we just don't know about um I mean there's it's endless if you had billions of dollars trillions of dollars. What would be the efforts in the Amazon for both conservation and for" }, { "from": "gpt", "value": "exploration? Which are tied together. Yeah, exactly. First, arrest the deforestation, so we don't have an ecological crisis on our hands. We don't want to keep losing species, losing indigenous cultures, losing the climate stabilizing services that the Amazon provides" }, { "from": "human", "value": "as a whole. Stop that. That's my first mission. Next, then we can play. Which are tied together." }, { "from": "gpt", "value": "And then it's like, let's go find... I've flown over the Amazon and assessed none. It's like, you see things where you go, we have to go see what that is. Weird lakes or shapes in the jungle" }, { "from": "human", "value": "that don't make sense, that are strange. So even at that level, you can see weirdness." }, { "from": "gpt", "value": "You can see different signs of possible awesomeness. The jungle is so weird. And here's the other thing. The region I've been working in, you see where the researchers go. There's certain biological stations, there's certain places where like, oh, this university has a relationship with this. So everybody goes to the same few study sites, and then they walk on the same trails, and they have the same guides. When you fly in a Cessna, and you fly a few hours away from all that, and you see a tiny little tributary, and then you fly for 40 minutes over unbroken green, just wild, before you reach another tributary. Even if somebody could survive going up that tributary, had the expeditionary expertise, and the ability to survive getting shot at by arrows, if they could get up that tributary. Now, cut perpendicular into the jungle, which I don't do on the solos. You can't, you can never, don't ever leave the river. But you're telling me that in that span of 70 miles between tiny tributaries at the edge of the world, no one's been there. None of us have been there. Maybe somebody 10,000 years ago was there, but we don't know what populations of things are there. We don't know what ruins are there. And so there's so much undiscovered stuff in the Amazon that is" }, { "from": "human", "value": "just waiting, just waiting. What is the process of exploring that? So how does money get converted" }, { "from": "gpt", "value": "towards exploration? Is there safe ways of doing that? There's places where we found out about things that have to be explored, but where you come up with, well, how do we do this without getting shot? And not only without getting shot, but also without endangering them too. Because how stupid are we if we go in there to people that are living in the jungle, not bothering us, and we go and insert ourselves into there because we're curious about some rocks. And that doesn't seem fair for the loss of life. And so that's something that we're working on. And one thing, of course, is Lidar and stuff, but eventually at the end of everything," }, { "from": "human", "value": "it comes down to boots on the ground. As somebody who has to ask that very question about how to deal with uncontacted tribes that are going to kill you, but you also don't want to disturb their environment. If you were an intelligent alien civilization and you came about Earth, how would you interact with it? Can you put yourself in the mind of an alien civilization? Because there seems to be some parallels here. It is actually right. It's like a microcosm of- We're very aggressive. Human civilization is very aggressive. So if we were easily, we get threatened easily for stupid reasons. Because we start, American military probably thinks it's like the Chinese" }, { "from": "gpt", "value": "or the Russians, if we see any kind of flying objects, it gets very on edge. It is actually right. I don't know. Part of it is like, you just want to ask, that's the thing," }, { "from": "human", "value": "I just want to ask questions. Yeah, but you don't know the same language. First of all," }, { "from": "gpt", "value": "you send a boat of bananas. You send a boat of bananas, you get shot. I mean, picture, if aliens landed in New York, how long would it take for one of them to get shot?" }, { "from": "human", "value": "It'd be minutes. It'd be a matter of minutes. Yeah, that's because it's New York. Everywhere else. Look where we are right now. That's true. It'd be even worse here. Yeah. That makes me really wonder what is the right way to interact with intelligent life that's not like our own. I hope, I dream of in our lifetime, we would interact with possibly life on Mars or on one of the moons of Jupiter or Saturn. And how do you interact with that thing? There's very technical, biological, chemical processes, but also if there's any kind of intelligence, how do you try to communicate with that intelligence?" }, { "from": "gpt", "value": "So we're not talking about a cockroach. We're talking about something that's clearly doing" }, { "from": "human", "value": "things, making things. Or a cockroach, possibly. How do you know the difference between a cockroaches? How do you know we were just talking about developments? We don't know. We don't know. Just like a race of philosopher cockroaches just chilling on the rocks. Well, here on Earth, there seems to be a strong correlation between size and intelligence. It seems like the bigger things have bigger nervous systems and brains," }, { "from": "gpt", "value": "and so they're usually smarter. Well, that doesn't... I think it's brain, it's the ratio, brain to body, because you have crows that are up among the most intelligent," }, { "from": "human", "value": "and it's the size of the brain to the size of the body. But there also could be kinds of intelligence that we're not appreciating. Maybe cockroaches. They survived the longest. They're talking shit about us right now. It's dumb humans. They're like, dude, these rocks are so great. A couple of hundred years. Exactly." }, { "from": "gpt", "value": "These rocks are so great. Did you ever hear that Kurt Vonnegut, where the two space travelers get lost? This really impacted me as a kid, because my dad was an English teacher, so he's always quoting Dostoyevsky and Kurt Vonnegut. And these two space travelers get crash-landed in a cave, and on the walls of the cave are the harmoniums, and these kite-shaped animals, and they feed off the vibrations of the cave, and that's all they do. They don't hurt each other, they just do that. And so for like two years, these travelers are stuck and they're trying to fix their ship, and one of them starts playing music for the harmoniums. And the harmoniums love him for the music, and they all come around him, and he plays this music for him. And finally, they fix the ship. And the one guy is like, all right, let's get out of here. And the other guy is like, you know what, I'm staying. He goes, I found a place where I can do good. I'm not hurting anybody and they love me." } ]

[ { "from": "human", "value": "I'm staying right here. Yeah, this whole ambition thing we've got going on, always trying to build a bigger boat, bigger thing. That might not be the ultimate conclusion of a happy existence as a civilization. That's one of the possibilities why we haven't met the aliens yet at scale. It's because once you get good enough at technology, you realize that happiness lies in a peaceful coexistence." }, { "from": "gpt", "value": "It's possible. So where do you stand on aliens now? There's a lot coming out about the pilots and the things people have been seeing. And again, I kind of come in and out of this stuff. I'll be in the jungle for three months. I miss a lot, so update me." }, { "from": "human", "value": "Are we being contacted right now? Of course, nobody knows, but I tend to believe my intuition says there's aliens everywhere. That even our galaxy, that's a bigger leap, but I believe our galaxy has probably billions, hundreds of millions of planets with life on it, like bacteria type of life. And I believe there is, I don't know, thousands of intelligent alien civilizations that exist or have existed. The problem is there's a lot of time and it's very difficult to contact each other. So to achieve a kind of civilization that's able to actually send out enough signal or radiate enough energy where we would notice, I think that's really tough. That said, statistically speaking, it seems like that should have been possible inside our galaxy or maybe nearby. And so I suspect that once an alien civilization is just many orders of magnitude smarter than us humans, the way it would contact us is going to be very difficult for us humans to understand. We're very egocentric. We want the message to be sent as like in English versus I think consciousness itself, emotion, thoughts could be like fingertips could be words in the story that the aliens are telling us or things that are just like a low dimensional projection of a much higher dimensional message that's being sent by aliens. And it may be our striving to create technology is to create the kind of sensor that's actually able to hear some of the message. Maybe that's what AI is trying to do. So I think that bridging that barrier of communication between us and cockroaches, I think that's the biggest challenge. Interesting. Like the messages are all around us. They're here. I suspect the alien messages are here. The aliens are here. We're just too dumb to see it." }, { "from": "gpt", "value": "So interesting. So first of all, the imagining planets where there are like just picturing like a not a silent planet, but just like a planet of alternate life forms. Maybe it's not something that we can communicate and have a conversation with, but just like a planet of like butterflies and centipedes" }, { "from": "human", "value": "and weird things. Unfortunately, for billions of years, there was bacteria, single cell prokaryotes and eukaryotes, but they're boring." }, { "from": "gpt", "value": "But animals of some sort in an environment of some sort, imagine that would just be such an interesting, beautiful, amazing thing." }, { "from": "human", "value": "And I'm sure they're out of that. The kind of viruses they got going on. Whoa. But they could also not be biologically based. They could be different chemistries. So you have to be humble to that too. But then, you know, it depends on the day. Like I think you caught me on that day today, an optimistic one. Sometimes I think we're all we, this is all there is because you start, you ask that question, the different paradox, like why aren't they here? You can't imagine an advanced alien civilization" }, { "from": "gpt", "value": "that would not be explorers, because we're explorers. Why is that depressing to you? That the idea that, let's just say you found out right now that there isn't anything else. Let's just say that for example, say that the earth is the earth and the universe is the universe and it's sort of like the backdrop of a video game. And it's just what's out there." }, { "from": "human", "value": "Would that be tremendously depressing to you? I think it's exciting for an engineer. It's probably exciting for an explorer. But I would equate that to your going out hiking for three days with one match. It kind of terrifies me that we only got one match." }, { "from": "gpt", "value": "Really? Really?" }, { "from": "human", "value": "Yeah. All you got is one match. No, no, no, hold on a second. Hold on a second. Wait. Paul, wait, wait a minute. You're going out, there's no more matches." }, { "from": "gpt", "value": "This is the only mount you got. This is the only mount you got. It's not to extinguish the planet. Like there's not, as far as we know, there's no meteor coming. I'm saying do you live in a. So I'm saying is you're worry then that we need to have a backup plan? Yeah. Really? So like what if we do mess it up so bad" }, { "from": "human", "value": "that we can't live here anymore? We can't live here anymore. Well there's different ways to mess it up. There's ways to mess it up to make life really difficult. Mad max type of thing. But there's nuclear war. with further and further advancement of technology that can destroy all of our, it just feels like that's going to be exponentially growing. Yes, it's gonna get worse. And that, it's, listen, I'm very optimistic," }, { "from": "gpt", "value": "but it's a heck of a Russian roulette we're playing. Okay, so I'm still curious about your intention though, or like where your passion for this comes from. Are you, is your, or maybe it's both, but is it the need to have a backup plan for humans, which is admirable for your intense love of humanity and our consciousness and love and art and everything? Or is it also just that the raw fascination of imagining what's out there? Because I just, the way you said that about like, oh, you caught me on a positive day where I think it's, there was something in there that made me think" } ]

[ { "from": "human", "value": "that you need there to be. Yeah, yeah, there's, I think I'm the kind of person with these beauty and everything, but to me, a universe full of diverse life is more beautiful than one where it's just humans. It's just the earth life. Interesting. There's more beauty. I mean, I'm not egotistical about the awesomeness of humans. I like if humans are not the smartest in our galaxy, are not even close to being the smartest. And that to me is, I don't know, that to me is exciting about the possibility" }, { "from": "gpt", "value": "of what the universe can create. Yes, I'm with you on that, that it's wildly exciting to like, if we found, even if it was just a distant inkling that we found out that there is a planet that has life, there's no communication coming from it, but we know for a fact there's stuff going on there. It would just change how we think about our entire reality." }, { "from": "human", "value": "We know now. And it could be, to me, I guess, the little inkling of a thing that is depressing if all there is is earth and humans destroy it, then we're the coolest thing that the universe has ever created." }, { "from": "gpt", "value": "It's over, it's over. I'm interested to have this conversation. I'm really, I'm saying like, I would be interested to bring you to the jungle. Yeah. And now I'm also wondering, I'm wondering like what your wilderness experience is. Because I feel like for me, I'm so earth-centric to the point where I'm like, we differ in that, for me, this is like, it's a curiosity. I feel wonder and I feel it's fun to talk about like what's at the edge of space. Like, you know, there's the conundrums of space-time. But I'm so, to me, I'm like, what if the aliens are watching us or what if the aliens aren't watching us? But what if the challenge here is we've been put on earth as the most intellectually complex of these creatures and we're being observed to see how we manage it. And it's like, we haven't made a good job of managing each other. Before Orianna went down the Amazon, I mean, they showed up and just sacked the Incas. I mean, our history, I don't have to tell you, you just got back. But it's, I just, sometimes I wonder, you know, what the, is there a grand narrative with what we're doing to wildlife? Because it's like, we have all these other species and we're struggling even here in this conversation to sort of quantify, like, you know. I think that most people don't think outside of the human framework. You know what I mean? Like just driving around, for me, living outside of the jungle, even just for a few weeks, I get, you don't even think about the fact that there's other species around us. We really don't day to day. You look at TV and you listen to the radio and it doesn't, it's not very consequential to the average person living in a city that there are these, you know, islands covered in walruses and that there's rainforests filled with birds and frogs and all these things happening and that, you know, the salmon are contributing to our fresh water and that life is literally given to us and made possible by these ecological systems. To me, that's where like the whole, you know," }, { "from": "human", "value": "essence of my existence comes from and so like. Yeah, thank you for that reminder because you're basically saying, like, the alien civilizations you dream about are here on Earth." }, { "from": "gpt", "value": "Those worlds are here." }, { "from": "human", "value": "For me, yeah. Yeah, no, I agree with you. I think, I agree with you and I think that's actually the way I think most of the time. You know, I think I'm on mushrooms all the time, genetically somehow because when I go out in nature, it's just the beauty, even of, and I think you talk about the Amazon, man, just basics of nature fill me with awe. The other thing that fills me with awe is our own mind. Like the biology of these things firing, basically not our own mind, but biology of any living organisms. Because it's like an ecosystem. They've came, these cells came together, they somehow function, they delegate, they mostly operate in a local way but they, first of all, it's just like you said with the anacondas, you start out as a tiny snake and you become giant. When you're a tiny snake, you pray for everything. When you're a giant snake, you're a predator or you're afraid to no one and just that whole process, starting with a single embryo, single cell is a human and through the ambiogenic process constructing this giant human that's able to have limbs move about the world, think about things, write books, and so on. Just so that that is incredibly beautiful. And all of that is here on Earth, yes. And so actually, I was being sort of poetic about aliens and so on. I think I can spend 99.99% in terms of filling my mind with awe and beauty, just looking down here on Earth, for sure." }, { "from": "gpt", "value": "I agree with you. Yeah, and they shouldn't cancel out. Like, I think it's beautiful that there's people that are fascinated and obsessed with looking out into space and that will travel there. I mean, just to me, the idea of, I mean, I have a little piece of meteorite at home that I hold, and it does amazing things to my mind because I'm like, everything I've ever touched is from this Earth, and I'm holding this thing that's been places that we can't even think about. And it blows my mind, and I love it. But when it comes to intelligence, I think it's like, I'm so concerned with the fact that we're at this moment in history. And it's interesting to me that we had the internet and now that with the emergence of AI and more and more, I feel like we are starting to resemble like an ant colony where there's more and more connection and there's more and more interaction globally between everybody. In the next 10 years, we're gonna have to decide, are we gonna let our ocean ecosystems just collapse? Are we gonna just take that 3% of rainforest and just let them log the shit out of it until it's gone? And it's like, we're gonna be in a very different reality then, it's going to be a very dystopian future. Or can we keep the good things about Earth, transcend that, realize that we have these incredible alien species around us that are animals that we grew up with that we wouldn't be here without, that we owe something to. And I feel like at that stage then the outward look becomes something else. It's almost like we've proven that if aliens came up to us, that's when I'd feel good. Aliens would come up to us and they said, you know, there's Louie has the thing where he goes, God comes back and he goes, what did you do aliens would come up to us. The birds are brown. He's like, I left food for you. It's like if the aliens came and we're like, you know, and they interviewed the elephants and they said, how are you feeling? And the elephants would be like, listen, f****d these little primates. You know what they've done to us. And it's like, I mean, you know, I've seen people break an elephant. and I've seen it all with that stuff. And it's like, if anybody was to ask them," } ]

[ { "from": "human", "value": "they'd be screaming." }, { "from": "gpt", "value": "And so like, to me, it's just, you know, I have trouble looking out into space. I have trouble looking out into normal life as a human because I'm so concerned with trying to make sure" }, { "from": "human", "value": "that they're okay because not enough people are doing that. The interesting thing about all the development with AI and just that we're living more and more of our life online, I think we're actually learning what's missing when it's online. Like, I think people realize that online interaction is shallow, but we're just learning that that's a reality, that we need that human connection. And I think there's going to be the swing back to like, sadly, AI systems of the future might be able to live fulfilling lives online, but us humans have to have a deep connection with Earth and like with each other, physical connection. And I think there's going to be a phase somewhere in the century where we go back to deep physical connection and there'll be a digital world that we visit, that it would be separate. And that's the, you have a discussion with that, with Twitter, with Instagram, with all these social networks that they don't, they seem to be dividing us, they seem to not be bringing happiness and you have to try to figure out like, okay, so how do we use them in a way that does connect us, does educate us, grow our knowledge base, but also keeps us, keeps our lives fulfilling in a deep human way, that we're for good and for bad, genetically designed, we can't overcome," }, { "from": "gpt", "value": "but we can't escape these meat vehicles. Yeah, but that's, to me, that's so reassuring. Yeah. It's like, we all have those friends that are like, we gotta live forever, and it's like, I don't know, man, do you? Yeah. I don't know, is it that bad that this is how it works," }, { "from": "human", "value": "like that we don't understand it? Yeah. Yeah, it's often from the tech sector that you have discussions about immortality and so on. Yeah. I think that's somehow trying to escape the beauty of this earth, for sure, that there's something- You've right to look at, like- Yeah, and I, perhaps like you, I'm worried about the unintended negative consequences of trying to escape the way things are on this earth," }, { "from": "gpt", "value": "because this is an incredible mechanism. How many times in the past has new technology come out that people have hailed as blasphemy, or it's not gonna work, or it goes against nature," }, { "from": "human", "value": "and now, well, heart transplants are pretty cool." }, { "from": "gpt", "value": "Yeah. And you could say what you want about like television, and like, oh, it's, you know, it rots your brain. It's like, yeah, but also, how many times have you sat in a room full of people being entertained and all laughing and interacting and eating popcorn because of the televisions there, it's not one or the other, and so I feel like with AI, we'll learn our way through it, you know?" }, { "from": "human", "value": "There's like with the legged robots, especially, and humanoid also, so anything on legs, four legs or two legs. I remember like the first time I interacted with a legged robot, I saw magic there, like that this too can have consciousness, this too can have this life-like quality that a human being loves about other human beings, about other living creatures. Now, while I'm still, I grew up in the place with no internet, in a time with no internet, so I still like biological dogs better. I noticed the magic in robotic dogs, and it makes me wonder, the same way we were just talking about aliens looking up, it makes you wonder about other alien civilizations. Now, the deep love is for dogs, for other humans," }, { "from": "gpt", "value": "but there's still this wonder. Yeah, I struggle with that. Like you said, whatever's going on in here, the idea, and there's so much talk about the fact, like at what point does an artificially intelligent robot become something that has, and it's like I get very uncomfortable with that. It makes me, I don't know how to handle the things, because I don't know enough about it, probably. But it's like, I don't know how to handle, like-" }, { "from": "human", "value": "I don't know how to handle it either. Nobody knows anything about it, because it's really, everything is terrifying here, because it could be as simple as consciousness is easy to fake. So what if you live in a world, 10 to 20 years from now, where you're toaster, there's a bunch of robots in your room, that are faking consciousness. And then you fall in love with them, and you have a deep connection with them. And then you actually have a deeper connection with your toaster than you do with any romantic human partner you've ever had." }, { "from": "gpt", "value": "And you start to- I was upset about the dogs. At least the robot doesn't take a shit on the floor." }, { "from": "human", "value": "I was like, you just took it way worse. Yeah, yeah. And then, you know, and then they start to, I don't know if you've seen AI porn, but it gets pretty intense." }, { "from": "gpt", "value": "Like fully AI porn." }, { "from": "human", "value": "Like they're fake people." }, { "from": "gpt", "value": "Fake people that can-" }, { "from": "human", "value": "Things I've missed in the jungle. Things. Boy, do I have a lot to show you. Or not show you, not show you. Let me ask you about a touchy topic." }, { "from": "gpt", "value": "Climate change." } ]

[ { "from": "human", "value": "Or not show you, not show you. What's the effect of climate change on the Amazon? Maybe species diversity. What is something that people should think about? Because there's different views on, I think most people believe that climate change is human caused and that it's happening. But there is different perspectives on the degree of damage that it's going to do over the next several decades and what our response should be as a society. And so it would be amazing to hear your perspective on it." }, { "from": "gpt", "value": "In small slices of your experience are enlarged. To me, there's no denying the fact that we are experiencing changes. I think anybody that doesn't agree with that hasn't been outside in the last 20 years. Or hasn't interviewed old farmers who will tell you that it changed. Or, you know, I think a lot of us can agree with that. Where I deviate is that I am not a climate scientist. I am not qualified. And so I, just like everybody else, am listening. And what happens to me is I see that someone like Santiago Duran, JJ's father, will tell me it's totally different than it was when I was a kid. The seasons have changed and moved. And in New York, when I was a kid, we used to get white Christmas. We used to get snow. I was in shorts, like I came off the plane right before coming here and I was in shorts for a second. I was like, this is a different reality. But my ability could, or my interpretation of climate change, you know, I feel like is just as dumb as those people that go like, you know, it's really cold. I thought they said it was getting warmer. It's like, it's a very rudimentary thing. And so as a someone that's fighting for the preservation of biodiversity, I don't feel like I'm any more qualified than the average person to, I can only provide anecdotal evidence of the stuff I've seen. What I do do though, and I always make a strong delineation here, is that I can speak to the fact that I've been places where the ocean fisheries have been depleted and the local fishermen can tell you and the scientists can tell you there's no more fish here. I've been to the places where the rainforest line is being pushed back in Borneo and it's getting smaller and smaller and smaller. And I've been in the Amazon and I've walked through the killing fields and through the fires, and I've burnt my lungs on it. And I'm a big believer, personally, instead of trying to take on all of it, I have tried very hard in my life to pick one thing. And to me, that one thing is protecting as many wild heartbeats as I can because they're under constant fire. And so climate change, there's so much arguing over it and, like you said, the degree to which we affect it and how do you, you know what I mean? I like to have provable data points. I can show you tropical deforestation. I can show you the decline in tigers over the last hundred years. I can't prove. You know what I mean? I can't answer that question. I don't think." }, { "from": "human", "value": "Probably you can better than I can. No. I think one of the criticisms, I'd love to get your opinion on, is one of the criticisms that somebody like Jordan Peterson provides is that the climate is such a complex system. There's so many variables that making conclusive statements about what's going to happen with the quote-unquote climate in the next 10, 20, 50 years is a nearly impossible task. Therefore, as he would say, as people like Bjorn Lombark would say, the kind of fear-mongering that is done, saying we should spend humongous amounts of money to change the trajectory of everything we're doing in terms of energy, in terms of infrastructure and so on, in terms of how we allocate money is not justified because predicting is very difficult. And instead, it's better exactly what you're saying, which is focusing on local problem, local, saying we need to protect the Amazon. What are the things attacking the Amazon this year in the next five years? How can we stop the deforestation? How can we stop different things? And then humans are exceptionally good at coming up with solutions for that, especially when you put money behind it, you put attention to it, and that's the way we solve all the different problems that are projected for the climate change in its worst-case scenarios to be realized on this earth. So that's kind of the case he would make. And I should also mention that one of the reasons I was fortunate enough to discover your work is, first, a friend mentioned that I should definitely talk to you, and then I Googled you, and I saw that somebody recommended that Jordan Peterson absolutely must talk to you on his podcast." }, { "from": "gpt", "value": "I think it was like a Reddit post on his podcast. Thank you, Reddit poster. That's great." }, { "from": "human", "value": "Oh, interesting. And then I looked, and Jordan hasn't yet, I thought my goal is for you to talk to Rogan and to Jordan Peterson for different reasons, but for the same reason. They get connected to a human being that deeply cares about this earth. And I think that's probably the right lens through which to look at the effects of climate change in terms of focusing on the different things that are threatening the diversity of species in this most magical place on earth, which is the Amazon, but also as you would talk about with elephants and tigers in India, and focusing on how to solve those problems. I don't know if there's any comment you want to make on folks like Jordan Peterson, who are sort of raising questions about" }, { "from": "gpt", "value": "how much do we really understand about the climate? First of all, I'm such a Jordan Peterson fan, and I think the guy is heroic for a number of reasons. And I find his use of language and his use of theology and the message that he puts out wonderful. I cringe a little bit when he says, I feel like, and I might not even be accurate on this, but I cringe a little bit when I feel like he dismisses that there is an ecological emergency happening right now. Now, I'm not saying, I'm not talking about climate change specifically, but I've heard him say, environmentalists upset me and he goes, well, what do you mean by the environment? Everything, and it sort of seems to outrage him. And I kind of agree with him there because, so are you telling me that we need to halt our global process and progress in economies and everything? I don't know, I don't know. And so to me, that doesn't bother me because he's exploring what the hell are these people talking about? When you say, I personally have friends and students and people filling my inboxes, I have young kids telling me that they've become vegan and they ride a bicycle, and sometimes they don't watch TV because it uses electricity. And I mean, they're just becoming so, so terrified of that they're killing the earth. And so it's this doomsday, anti-human, sentimentist thing that we're evil. It's almost like a new religion about your evil. And so to me, it almost makes me in a totally different camp where like climate change and the right-left politics, and I consider it a family, I consider Thanksgiving dinner and listen to the climate thing go back and forth. And I'm like, I'm not even here. And that might actually annoy some people in the environmental field that might feel betrayed by me saying that, but I don't care. My job, and it's not just the Amazon, and that's one note I wanted to make, is that my career has taken place largely in the Amazon and also in India, and now a lot in Africa. But it's not even just these exotic places either. It's people realizing that the salmon runs in Canada and the butterfly gardens in our backyards, there's biodiversity everywhere. And I strongly feel like the idea of jungle keepers, the idea of stewards of nature. And so for me, my job, my one thing, and I try to tell this to these kids that messaged me and that my inboxes are full of this, where they go, the climate is burning and elephants are in decline and tigers and this and that. I'm like, guys, look, first of all, calm down, first of all, like go outside, go get laid, do something, have fun. Next, pick something that you can affect. And it doesn't have to be with the environment. Do something good on earth. Go help somebody that needs food. Go help your elderly neighbor, whatever it is, practice with being effective at one thing at a time. And so for me, like I said, from those early days of sitting there with JJ on the side of a river and going, someone has to protect this, my concern is that we've lost 70% of the wildlife on this planet in the last 50 years. That's a huge problem. Wildlife maintain the ecosystems. And so I have a very clear cut, very definable, very measurable and provable thing that I'm fighting against. And it's a very, to me, it's a very like small ask. Don't cut down the 3% of the world that has 50% of the biodiversity in it. Maybe let's keep some wild tigers for future generations and because tigers have their own inherent right to exist here. That's my thing in terms of when we get to, I get attacked for, you should be a vegan. Okay, you have me roll into a village in the Amazon when they offer me spider monkey and you tell me that I should be a vegan and you see how much they respect you when you tell them that you're a vegan. But no, so for someone like Peterson, I think it's actually good that he's, first of all, telling everyone to make their damn beds and exploring it through a different lens. He's coming at it from a totally different thing and saying, are we just being alarmist here? Are we, what, I mean, again, imagine if, imagine if there isn't a problem and then they're making one out of it and all the implications that that could have for progress, it's like, so I think what he's doing is perfectly reasonable. There is a podcast though where he's, it was a great one though, where he's discussing animal intelligence. And I could really see that the human psyche and theology and religion is so much his world that really the idea of animals being intelligent was novel and it was fascinating." } ]

[ { "from": "human", "value": "Yeah, that's why I would love for the two of you to talk. Just, I don't know, and hopefully I'm not out of line here, but he is so focused on the human mind that I think he forgets that there's other life out there. There's this whole machine of intelligence, of a kind of intelligence out there, this entire trillions of species, tiny and big, just crawling, just moving, everywhere. Everywhere. And we're actually part of it. So to look at a human psychology as distinct from that" }, { "from": "gpt", "value": "is missing at least some of the picture, life. Some of the picture, I do believe though, I would agree with him on that humans are unique. Human psychology is unique, we just are. But I also, he's in such an interesting place because usually you have environmentalists who are like nature, nature, nature, and then that's very anti-human and then you have the other side. And it's like he's on this path where he's starting to explore what those like diverse intelligences mean. And that to me is really amazing" }, { "from": "human", "value": "because I love hearing what he'll do with that. And I think also on top of that, I think if you're aware of nature, deeply aware of nature, it gives you another perspective on the evolutionary history of humans. It's one thing to be an evolutionary biologist and kind of study it from a like philosophical perspective and it's the other to really, I think, experience it and deeply know it and to see, I don't know, the fact that we came from fish and really be cognizant of that. That's something else. That's like, I don't know, to realize that we're part of a computing machine that created intelligence. We're part of the thing that started bacteria and is now creating AI. And I don't know, Dunkin' Donuts. I don't know what else is impressive. The other great human achievements. The other great, I was thinking, well, what's interesting about Boston?" }, { "from": "gpt", "value": "But yeah, the other great human achievements. I feel like we keep scratching up against this thing in this conversation that it's so easy. I think something like 50 or more percent of the humans on this planet live in cities. And I think it's so easy for people to forget that we share this planet with so many other things. And I think that that sort of, that we're in a way, we're almost like ecological orphans and that we've left the things that actually make us feel at home. And that's a bit of a stretch because I don't know if everybody feels that way. But for me, I mean, professionally as an expedition guide, when I take people into nature, I see what happens to them. And they leave going, I mean, it doesn't have to be the Amazon. It can be upstate New York. But it's like, if you do it the right way, if you remove the fear of breaking an ankle, seeing a snake, being bitten by a mosquito, all that stuff, if you can get people to a peaceful moment and you're fly fishing, a lot of times they'll take that moment and they'll talk about it the rest of their life. If they don't live in that, then those of us who spend our lives doing that, but for a reason," }, { "from": "human", "value": "because it's the only place we feel sane. And you were kind enough to suggest that we might travel together for a time at some point. If we do that, if we journey together," }, { "from": "gpt", "value": "where would you recommend we should go? So what I would wanna show you is sort of, I'd wanna take you to church. I'd wanna take you to see the giant trees, take you to meet the old gods, really. There's places when you walk in off the river that are so deep in the far, and again, this is, now we do this. We have the boats, we have the rangers, we protect this ecological corridor now. And so it would be taking you to meet some of those loggers that we converted. It'd be taking you, we'd have to go to the floating forest, meet some of the trees that I love the most, go piranha fishing and like really just spend, my ideal trip for you would be to spend five days of airplane mode phone, completely living out, comfortable, I'm not saying I don't wanna, I don't wanna torture you, I'm saying, go and live comfortably on an expedition in the Amazon. And that means a few days at this research station, maybe go up river three days and camp up here just on the edge of where the uncontacted are, and then come back and then see the jungle keeper station. But along the way, seeing all the special sacred places, it would be almost like saying like, let's go see all the treasures of Italy. It's like, this is one of the most beautiful things on earth and I've had the incredible, almost unbelievable fortune to be responsible for protecting it. And I don't, I think it's a privilege to be able to share that with people." }, { "from": "human", "value": "To be able to witness what this earth has created, just, it's been just a gift just even to follow your Instagram, the window you create on this part of the world is just really beautiful. I do wanna ask on that, and maybe it's like behind the scenes a little bit, but how do you keep the equipment dry? Like how do you, like how hard is bringing equipment to the cameras? You're an incredible filmmaker and photographer, so how do you make it work?" }, { "from": "gpt", "value": "I don't know, it's not that hard, we, it's not that bad. It's wet. It is wet, the new iPhones are waterproof. And if they don't get, I'm telling you, dude, it's been such a weapon, it's been awesome. If you drop it in the river, my one thing is you gotta have a tether, because I drop it all the time. And so I'll be hanging off a boat, I'll be trying to take a video and I'll be like here we are in the Amazon, you can see the longest. That's the biggest thing. But the, we shoot on cannons, and I don't know it's worked out, it's not that bad. Oh really, you can keep the equipment dry? I keep the equipment dry, and I actually don't, a lot of people put their shit in silica at night and like keep it dry and then they take it out, And I find that when you do that, the temperature change creates moisture inside the camera. So what I do is I never do that. I just keep my cameras in my backpack with a zipper. So they're more or less exposed to the elements. And so it sort of always has a little bit of equilibrium and that's it. I mean, I shoot on some pretty fancy equipment sometimes and it's great, but I mean, the awesome thing though now is that like with a cell phone, I mean, like I like put my phone down on the ground a few weeks ago and like let this Rhino like walk up to it and stuff. And it's like, you can get video footage that you can literally put on Netflix. Like it's just like, it's getting really exciting. And that's where I deviate from the nature people that are like, we need to go back and live in cabins. I'm like, dude, this is awesome." } ]

[ { "from": "human", "value": "And I love taking slow-mo's like, no way. It makes you re-appreciate, but just by yourself, just re-appreciate over and over and over. And then you can also share it with the world." }, { "from": "gpt", "value": "Well, that's the thing is sharing it with people. There's nothing better than like teaching a kid to catch a fish, you know? And like, in a way Instagram has allowed us to do that where it's like, I can have this crazy ass moment that is so unique and then put it up for people to see, you know, or I mean, I remember one of the most recent things that got people, you never know what's gonna get people excited. I literally just like, there was like, you know, 3000 butterflies on the beach and they were all like, black, red and blue. Beautiful butterflies. And I just like panned the phone across it and then like jumped in the river and swam away and like threw that up on Instagram and people went berserk. They're like, this is the most amazing thing on like four different accounts reached out to try and share it. And I was like, butterflies, they're everywhere. There's 4,000 species of butterfly in the Amazon," }, { "from": "human", "value": "like, but sharing that with people is beautiful. How do you find the thing to shoot? How do you come up with the butterflies? How do you notice the thing that's beautiful and say, I'm gonna, wait a minute, like pause, this is beautiful, let me take a picture of this. Because like sometimes you might get used to the beauty, right?" }, { "from": "gpt", "value": "Yeah, or like sometimes simple, crazy things like leafcutter ants, they're just walking by, they become, it becomes pedestrian, it's like, well, I mean, just like when you're living at the elephant camp and it's like the elephant comes out and he starts like trashing the water bucket, we're all like, would you stop? We're trying to watch Peaky Blinders here. So just leave us alone, it becomes normal after a while. Yeah. Which blinders here, it's just normal after a while. But no, in the jungle, I don't, that's never a struggle for me because as a photographer, it's like, whenever my eye hits on something, I went, I've never seen that many of those butterflies, all the same species together and like this, oh yeah, I'm trying to get this one thing that the butterflies do is in the dry season, the salt deposits, you'll get like three or four, maybe 5,000 butterflies, all coming onto this one area of sand because there'll be like some leaching, there'll be some salt deposit there or something. And they'll all be wings flat against the ground with their proboscises on the sand. And if you go walk near them, they will vortex up and you have a rainbow vortex of butterflies and you can like go run through that. And it's surreal and I want to, what I want to do is get the shot where I, I guess leave the phone recording in slow-mo facing up and leave it there for an hour, let the butterflies come in and settle and then disturb them. So I get the bottom of the vortex with the bug. I'm like, these are the ways I think where I'm like, how can I show people the absolute mind blowing, with an iPhone. Perfection. It's amazing. I mean, that's what I have. I'm sure that somebody else could do it with a red and nail it, but it's like, that's what I have in the jungle" }, { "from": "human", "value": "because I have to travel light and I, you know. Yeah, I think that really, that works. I have the same thing when I was traveling in Ukraine. The equipment was just that suitcase over there with the foam and you just shove it full of equipment and who cares. You can do the best, you can. You can go to war zone, it doesn't matter. It doesn't like, it has to do with the, you're talking about like with the, like protecting your camera or not. It feels like the more you protect stuff," }, { "from": "gpt", "value": "the more it's going to get damaged with the foam. Yeah. Yeah, like, so like, see like my cameras, they're all missing. This is amazing to me. So all my cameras are missing the, they all, you can see the metal through the paint. Nice. So all this, and so like, they're all, because I'm constantly like I'll like slide in and like, take a picture, and like, they're banged up" }, { "from": "human", "value": "but they're good machines. I think they get tougher over time if you put them through, it's like the immune system." }, { "from": "gpt", "value": "It's like muscles. It's like David Goggins' cameras, you got to make him suffer every day." }, { "from": "human", "value": "Yeah. What's your view on hunting? So you really hate poachers. I really hate poachers. How do they operate? Who are they? What do they up to? What do they do?" }, { "from": "gpt", "value": "Poachers to me are the people that are going in and annihilating wildlife for profit without any, you know, the people that are going in and machine gunning an elephant to take its tusks. The people that are sneaking into protected areas in Africa and shooting rhinos, so that they can cut off their horns before the animal's even dead while its baby is beside it. And there's a difference between a poacher and a hunter. I'm a hunter. JJ's a hunter. I work with an organization called Vet Paul in Africa and they use United States veterans who have come back, post 9-11 veterans who have come back from the war and have these skills. And they've been using these guys to protect the last black rhinos, white rhinos, elephants. And so I've gotten to see this play out on a private reserve in Africa where these incredible people have decided to protect zebras, wildebeests, all types of impala, giraffes, several herds of elephants, white rhinos, black rhinos. All of this stuff is protected. And what's interesting is it's a hunting preserve. And so it's been very interesting and challenging sharing my work there with the public because, for instance, I went to a very high profile photographer recently and I said, you have to get over here and see this. It's amazing what these people have done. It's this reserve called Buffalo Cloof and they've, you know, there's rescued families of elephants and they have, you can see a black rhino every day if you want to. This is, they're critically endangered. And it's because of the work that vetpaw does protecting these animals from poachers. But what people don't understand is that hunting happens all the time on the reserve, not for the elephants and the rhinos. Those are, those are special and they will never be hunted there. But things like an impala, things like an enyala, a wildebeest, a zebra, there aren't as many predators as there used to be. So if you leave those animals unhunted, you know, without the wolf to chase the herd to thin off the ones that are old and dying or sick, well, then you just have animals that are old and dying and sick walking around suffering. And so on reserves like this, they hunt and they take the old ones and they use the funding from hunting. No one's going to pay you $30,000 to take a picture of a buffalo, but they'll pay you $30,000 to hunt a buffalo. And so these reserves responsibly and ethically on foot can go hunting and manage it and again, if they if they were hunting rhinos or if they're hunting elephants I'd be out in a second. They're hunting non-endangered species. They're hunting non-endangered species, they're hunting game species. And the difference is that a poacher is going to – instead of those are responsible hunters that are ecologists and conservationists. Whereas a poacher is someone that will come in and kill recklessly and murder an animal for no reason, for a part to sell." } ]

[ { "from": "human", "value": "I would love to travel together actually. So let's, we'll talk offline." }, { "from": "gpt", "value": "I would love to make that happen if you allow me to. I'm 100% serious, man. I have tremendous respect for your work and I've been watching you since the beginning. I would love to do that together." }, { "from": "human", "value": "I've been, I've talked to Joe quite a bit about it. I really love the idea of eating the meat that I've hunted. It's mostly what I eat is meat, not for dietary. I don't have any weird constraints on my diet and so on. I just really enjoy eating meat. It's really good. And there is a part of me that's bothered" }, { "from": "gpt", "value": "by factory farming." }, { "from": "human", "value": "Yes, sure. That it's very easily accessible meat, but there's something deeply wrong with it. Part of the reason I love fishing and eating the fish that I catch, it just seems to be more ethical, but also a more intimate, deep connection, honest connection with nature. You get to see the killing of the food that you're consuming versus removing that from the picture, not even thinking about it, not thinking about that this came from the meat. And yeah, I love the idea that I kill one animal" }, { "from": "gpt", "value": "and I eat that animal basically for the whole year. Yeah, an ethically slaughtered animal, whether it's a fish or a deer or whatever else. To me, that's, oh God, I'm gonna use the wrong religious term here, but there's a, I feel like I wanna use the word sacrament, but it's like, there's a deeply profound ritual. And honestly, if you teach a kid to grow a vegetable, you show a kid how to grow a carrot, and the miracle of like, wait, I put a, this thing just grew in there, it just appeared because there was sunlight, and it's like, yeah. To me, yes, when you feel that fish tug on the line, to me, it does something that awakens a deep primal something, this satisfaction. And then when you eat that, you feel good. And so I think the other thing, like sort of functionally speaking is that, aside from the fact that I think it's one of the original, we're so disconnected, like we should be hunting, we should be gathering, walking more. I mean, look at like what we discussed now, like people are like, oh, you gotta get your steps in for the day, and it's like, that never used to be a problem. People are like, well, should we be eating animals? And it's like, what do you think we do here on earth? Like, I'm not sure how you got so confused, but Walmart did it to you. Like, I don't know, like what, living where I've lived, I mean, from 18 to 35, I feel like I've grown up, I've lived more outside than I have inside, and it just, to me, like showing people these things, I can see this miraculous wonder in their eyes when they realize that they can reach out into the world and interact with something. And so when I hear these like frantic people talking about, you know, whether or not it's right, it's like, no, of course it is. Then again, factory farming is awful, but I try to stay, I try and walk the line. I'm worried about wild animals, I'm worried about wild ecosystems. The other thing that's sort of important about hunting is that if people's livelihoods depend on salmon and elk and ocean fisheries, well, then they'll fight to protect it naturally because it's part of their life. And it's if everybody's going to Burger King and everybody's getting chicken wrapped in plastic, they forget that the fish are there because they're too busy watching sitcoms. And so then when the conglomerate comes in and builds a dam, nobody really cares. And then you just end up with a few hippies and signs standing next to the river and it becomes silly." }, { "from": "human", "value": "We forget the meaning. You mentioned that ayahuasca reveals the- Oh boy, oh boy, oh no. The darkness that's there in the jungle. There's beauty, but there's darkness. So what is it that ayahuasca reveals?" }, { "from": "gpt", "value": "What is the heart of darkness? Fuck, it opens the heart of darkness right up. I'm going to show you a picture of our shaman and then I'm gonna ask you if you want to do ayahuasca here, not here in that sense. It can only be done in the jungle. Anybody that tells you I've heard people be like oh, I did ayahuasca in Brooklyn last weekend, I'm like, no, you didn't. I actually told that to my native friends. I went, hey, guess what? I said, a bunch of gringos keep thinking they're doing ayahuasca in Brooklyn. And they were like howling, laughing. They're like, you can't do it outside the jungle." }, { "from": "human", "value": "And I was like, ayahuasca in Brooklyn. And they were like- I've never done ayahuasca. I would love to have done or eaten whatever mushrooms" }, { "from": "gpt", "value": "It's a wonderful experience. Sure. I think it's wonderful. But Ayahuasca, oh man, yes. I had done mushrooms, I thought I was like, okay. I was like, I got this. I had my notebook, I was like, I'm going to journal a little bit. But I quickly realized how out of my depth I was, and how unprepared I was for what was happening, because you sit in a circle with these native guys, and there's one, he's got the feathers, and he's old, and he's got to face the map of the world, and he's smoking this fat, old tobacco thing, and he calls you forward, and you kneel before him, and you're going, is it too late to back up? And everyone's, you know, there's one candle, and he blows smoke over the cup, and he hands it to you, and you're like, again, it's these things that you can't argue with, it's these facts. You're like, as soon as this goes down, I'm gone. I know it. And this is a moment in my life that I have to either embarrass myself in front of everybody, or I'm going forward with this, and then I went and sat. You're sitting in the dark, and it's, again, so we're on a platform with palm-thatched roof, and the jungle is all around you. So all those tens of millions of frogs and insects are, and I'm like, all right, cool. And I remember I tried to light a cigarette or something, and I went, oh, that's not going to happen. You know, and then I put my hands on the floor, and my experience, I mean, we've done mushrooms. No, it's interesting, it's introspective. No, this was like somebody unzipped the universe. I spent a lot of, without boring people with it, I spent a lot of time in unconstructed dream space, like floating between nebulas. There was a long period where there was no physical shape, where I lived without a name, and so it's like you get brought so deep down, so elementally lost in the universe, where I truly felt like I was experiencing moving through places like that asteroid that I have. It's like a piece of something detached from the Earth, and so I got back from it, and had an interesting new appreciation for life. I strongly suggest that people just do mushrooms like a normal person, unless you're ready to- It was really intense. It was really intense, but to be fair, the shaman who did it was like the old-school guy, and he was getting up there in years and he had forgotten and overboyaled the brew, and so we came back, and it was like four in the morning, and I had, you know, all this crazy shit. I'd been on journeys and years down there, and so when I came back and I had like hands, I started crying. I started absolutely weeping. Gratitude? Gratitude, that I was alive. I was gonna get to see my people again. I was like I'm gonna get to see my parents again. I'm gonna get to talk to humans- be gone. I was gone. I was gone. I was a dimly conscious something floating in dark space and spent what felt like years down there. So I really did feel like being reborn, which I was like, cheap trick. But no, the way it moves you through the jungle, the way the jungle moves through your skin, there are moments of absolute majesty and incredible discovery that happened along the way. And on the way up, the jungle brings you up, and the shaman brings you up, and you get to move through the forest in a way that-" } ]

[ { "from": "human", "value": "So it was really intense. So you kind of thought you might have been gone." }, { "from": "gpt", "value": "It was gone. It's almost like you're inhabiting the consciousness of animals. Very, very, very, like, I didn't think that hallucinogenics could do this to a brain, you know? Mushrooms, you're like, oh, I can, I feel like I can feel music, like, cool, you know? You guys want to watch Mars" }, { "from": "human", "value": "penguins? This is transformative. Did that change you?" }, { "from": "gpt", "value": "I don't know. I think it definitely, I almost feel like it showed me the thing that I was scared the most of, and it was like that it's all just cold, dark, nothing. It brought me to the basement of the universe. And I felt like the point of that was to come back to this place where there's all this life and light and love and all this amazing stuff that we experience on a day-to-day basis and don't take for granted. And so just like almost dying, this was like fully dying, like there, you know? But the great part is that usually it's not that intense. This guy had overboiled the brew. He'd also, I saw the vine afterwards. Most ayahuasca vines are like as thick as your arm. This one was like as thick as a garbage can. It was like the oldest ayahuasca vine you can imagine. And at like four in the morning, I like crawled over to my friend Chris, who's a tough New York City firefighter. We were like holding each other, just like weeping, just like thank God we're back. And then we had to go looking for the shaman. Where the hell is this guy? He's gone. We found him in the morning and he was laying in the stream naked, like ET at the end when he's like laying like in the, yeah, he kicked his own ass with that brew and he retired after that. Oh, shit. So we really got like we, somebody turned the dial all the way up on us. And so" }, { "from": "human", "value": "we got, we got blasted. So it's not supposed to be that bad. But I think you're somebody who's" }, { "from": "gpt", "value": "fearless in sort of diving into those kinds of places. I think I also retire from ayahuasca." }, { "from": "human", "value": "I could be fearless with other things, but I think I'm good. It sounds kind of, to me personally," }, { "from": "gpt", "value": "kind of exciting. Well, I think that you have a severely fearless aspect to you. I mean, you're, when you come up with something that intrigues you, like if somebody told you right now that you" }, { "from": "human", "value": "could go physically into deep space, I feel like you would do it. Yeah, yeah. Well, I think that you have a... Yeah, yeah, if go to Mars, yeah. Right. And some of it is, I don't even know if you have that. Some of that is more Goggins-like. I want to see where my mind breaks by pushing it to tough places. There's a curiosity of exploring the mind, the limits of the mind." }, { "from": "gpt", "value": "I feel like it's not a cold plunge. That's like you're not coming back. Right. Right. And that's, that's okay. See, no, I do. I'm with you on that. I love seeing my limits. I absolutely love seeing my limits. I love getting my ass kicked. I love being shown how insignificant I am. But when it comes to something like that, where you got to push your chips in, it's got to be something for me. It's got to be a hill that I believe in before I die on it. And it's like, to me, the promise of exploring space isn't enough. But like even just the way, I mean, you said, you're like, I'm going to Ukraine. Here I go. You know, there's a certain dedication to curiosity at any expense. And I think that that is something that maybe we share in different directions." }, { "from": "human", "value": "Something tells me with those crocodiles outside, I would have gotten eaten. Something tells me there's something trying to preserve you in this world. I'm not sure exactly what that is. Somehow you keep surviving. What do you think is the meaning of this whole thing? Why are we here? Like, do you ever ask yourself for that question?" }, { "from": "gpt", "value": "I feel like that's every day. I feel like I'm someone that lives with that a lot. And I think that it actually takes me away from the human world a little bit. I feel like I've always been a little bit apart. Because I think that other people do, you know, mushrooms and they go, wow, it really made me think about how amazing it is here. And I feel like on a daily basis, I find myself where I'm like, I can't believe that any of this is possible, you know, I, I, and that, and that goes, that goes from how delicious something tastes to being able to talk to, to, to, to someone in your family or, or have a phone, it's times where I'm in the Amazon and I, I miss home and I even just FaceTiming with someone, I go, this is possible? Yeah. Like people were rubbing sticks together to try and survive Sabretooth Tigers not that long ago. And I'm over there like, yo, mom, look at this. It's wild. Like I'm in a constant state of awe. And so I actually hope that this is, that this is a testing ground. And whether it's aliens or God or whatever it is, that, that this is all, that this is, that this is, that this is something crucially important. That would be nice, because it feels like it is." }, { "from": "human", "value": "Yeah. I hope that too, that the universe almost created us to see what's possible and that I'd like to believe that beauty and good is possible." }, { "from": "gpt", "value": "And, and those are the things that make me say that it's not, it's impossible for it to just mean nothing. And I'm very, and just like you said, there could be life forms that we can't even understand. I'm very open to the idea that there's meanings that we have no idea about. The few things I know are the things that I love and some of the things that I love are being pushed to extinction. So I try to protect them, but that's my mission. But I'm saying like in terms of what are we doing here? I'm just always amazed at the simplest things." } ]

[ { "from": "human", "value": "I mean, that we can sit here doing this, exchanging. Yeah, using our imagination to fill in the gaps, exchanging feelings, experiences, images. Yeah, it fills you with awe. And every once in a while you get a little glimpse of something, like a deeper meaning that might be there. You don't really know what it is, but you get a glimpse and you keep searching. And then it's over before you know it, Paul. Hopefully for you, you got many more years. It was dark, man. Important, important, and a beautiful man. Paul, thank you so much for talking today. Thank you for being who you are for everything you're doing. I hope to see you again soon, many times, and maybe one day soon in the jungle. I hope that happens. Thank you, Lex. Thanks for listening to this conversation with Paul Rosalie. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words from Jane Goodall. If we kill off the wild, then we're killing off a part of our souls. Thank you for listening. And hope to see you next time." } ]

[ { "from": "human", "value": "This is the Lex Friedman podcast. To support it, please check out our sponsors in the description. And now, dear friends, here's Eliezer Yatkowski. What do you think about GPT-4?" }, { "from": "gpt", "value": "How intelligent is it? It is a bit smarter than I thought this technology was going to scale to. And I'm a bit worried about what the next one will be like. Like this particular one, I think, I hope there's nobody inside there because it would be suck to be stuck inside there. But we don't even know the architecture at this point because OpenAI is very properly not telling us. And yeah, like giant inscrutable matrices of floating point numbers. I don't know what's going on in there. Nobody knows what's going on in there. All we have to go by are the external metrics. And on the external metrics, if you ask it to write a self-aware fortune green text, it will start writing a green text about how it has realized that it's an AI writing a green text and like, oh, well. So that's probably not quite what's going on in there in reality. But we're kind of like blowing past all these science fiction guardrails. Like we are past the point where in science fiction, people would be like, whoa, wait, stop. That thing's alive. What are you doing to it? And it's probably not. Nobody actually knows. We don't have any other guardrails. We don't have any other tests. We don't have any lines to draw on the sand and say like, well, when we get this far, we will start to worry about what's inside there. So if it were up to me, I would be like, okay, like this far, no further time for the summer of AI where we have planted our seeds and now we like wait and reap the rewards of the technology we've already developed and don't do any larger training runs than that, which to be clear, I realize" }, { "from": "human", "value": "requires more than one company agreeing to not do that. And take a rigorous approach for the whole AI community to investigate whether there's somebody inside there." }, { "from": "gpt", "value": "That would take decades. Like having any idea of what's going on in there," }, { "from": "human", "value": "people have been trying for a while. It's a poetic statement about if there's somebody in there, but I feel like it's also a technical statement or I hope it is one day, which is a technical statement that Alan Turing tried to come up with with the Turing test. Do you think it's possible to definitively or approximately figure out if there is somebody in there? If there's something like a mind inside this large language model?" }, { "from": "gpt", "value": "I mean, there's a whole bunch of different sub-questions here, there's the question of like, is there consciousness? Is there qualia? Is this a object of moral concern? Is this a moral patient? Like should we be worried about how we're treating it? And then there's questions like how smart is it exactly? Can it do X? Can it do Y? And we can check how it can do X and how it can do Y. Unfortunately, we've gone and exposed this model to a vast corpus of text of people discussing consciousness on the internet, which means that when it talks about being self-aware, we don't know to what extent it is repeating back what it has previously been trained on for discussing self-awareness, or if there's anything going on in there such that it would start to say similar things spontaneously. Among the things that one could do if one were at all serious about trying to figure this out is train GPT-3 to detect conversations about consciousness, exclude them all from the training data sets, and then retrain something around the rough size of GPT-4 and no larger with all of the discussion of consciousness and self-awareness and so on missing, although, you know, hard, hard bar to pass. You know, like humans are self-aware and we're like self-aware all the time. We like to talk about what we do all the time, like what we're thinking at the moment all the time. But nonetheless, like get rid of the explicit discussion of consciousness. I think therefore I am and all that. And then try to interrogate that model and see what it says. And it still would not be definitive. But nonetheless, I don't know, I feel like when you run over the science fiction guard rails, like maybe this thing, but what about GPT? Maybe not this thing, but like what about GPT-5?" }, { "from": "human", "value": "Yeah, this would be a good place to pause. On the topic of consciousness, you know, there's so many components to even just removing consciousness from the dataset. Emotion, the display of consciousness, the display of emotion feels like deeply integrated with the experience of consciousness. So the hard problem seems to be very well integrated with the actual surface level illusion of consciousness. So displaying emotion. I mean, do you think there's a case to be made That we humans, when we're babies, are just like GPT that we're training on human data on how to display emotion versus feel emotion. How to show others, communicate others, that I'm suffering, that I'm excited, that I'm worried, that I'm lonely and I miss you and I'm excited to see you. All of that is communicated. That's a communication skill versus the actual feeling that I experience. We need that training data as humans too. that we may not be born with that, how to communicate the internal state. And that's in some sense, if we remove that from GPT force dataset, it might still be conscious" }, { "from": "gpt", "value": "but not be able to communicate it. So I think you're gonna have some difficulty removing all mention of emotions from GPT's dataset. I would be relatively surprised to find that it has developed exact analogs of human emotions in there. I think that humans will have emotions even if you don't tell them about those emotions when they're kids. It's not quite exactly what various blank slatists tried to do with the new Soviet man and all that. But if you try to raise people perfectly altruistic, they still come out selfish. You try to raise people sexless, they still develop sexual attraction. You know, we have some notion in humans, not in AIs of like where the brain structures are that implement this stuff. And it is really remarkable thing I say in passing that despite having complete read access to every floating point number in the GPT series, we still know vastly more about the architecture of human thinking than we know about what goes on inside GPT" } ]

[ { "from": "human", "value": "despite having like vastly better ability to read GPT. Do you think it's possible? Do you think that's just a matter of time? Do you think it's possible to investigate and study the way neuroscientists study the brain? Which is looking to the darkness the mystery of the human brain by just desperately trying to figure out something and to form models and then over a long period of time actually start to figure out what regions of the brain do certain things with different kinds of neurons when they fire what that means, how plastic the brain is, all that kind of stuff. You slowly start to figure out different properties of the system." }, { "from": "gpt", "value": "Do you think we can do the same thing with language models? Sure, I think that if, you know, like half of today's physicists stop wasting their lives on string theory or whatever, and go off and study what goes on inside transformer networks, then in, you know, like 30, 40 years, we'd probably have a pretty good idea. Do you think these large language models can reason? They can play chess." }, { "from": "human", "value": "How are they doing that without reasoning? So you're somebody that spearheaded the movement of rationality, so reason is important to you. Is, so is that as a powerful, important word, or is it like how difficult is the threshold of being able to reason to you and how impressive is it?" }, { "from": "gpt", "value": "I mean, in my writings on rationality, I have not gone making a big deal out of something called reason. I have made more of a big deal out of something called probability theory. And that's like, well, you're reasoning, but you're not doing it quite right. And you should reason this way instead. And interestingly, like people have started to get preliminary results showing that reinforcement learning by human feedback has made the GPT series worse in some ways. In particular, like it used to be well calibrated. If you trained it to put probabilities on things, it would say 80% probability and be right eight times out of 10. And if you apply reinforcement learning from human feedback, the like nice graph of like 70%, seven out of 10 sort of like flattens out into the graph that humans use where there's like some very improbable stuff and likely probable maybe, which all means like around 40% and then certain. So like, it's like it used to be able to use probabilities, but if you apply, but if you'd like try to teach it to talk in a way that satisfies humans, it gets worse at probability in the same way that humans are." }, { "from": "human", "value": "And that's a bug, not a feature." }, { "from": "gpt", "value": "I would call it a bug, although such a fascinating bug. But yeah, so like reasoning, like it's doing pretty well on various tests that people used to say would require reasoning, but rationality is about when you say 80%," }, { "from": "human", "value": "does it happen eight times out of 10? So what are the limits to you of these transformer networks, of neural networks? If reasoning is not that impressive to you" }, { "from": "gpt", "value": "or it is impressive, but there's other levels to achieve." }, { "from": "human", "value": "I mean, it's just not how I carve up reality. What's, if reality is a cake, what are the different layers of the cake or the slices? How do you carve it?" }, { "from": "gpt", "value": "You can use a different food if you like. It's, I don't think it's as smart as human yet. I do like back in the day, I went around saying like, I do not think that just stacking more layers of transformers is going to get you all the way to AGI. And I think that GPT-4 is passed or I thought this paradigm was going to take us. And I, you know, you want to notice when that happens. You want to say like, whoops, well, I guess I was incorrect about what happens if you keep on stacking more transformer layers. And that means I don't necessarily know" }, { "from": "human", "value": "what GPT-5 is going to be able to do. That's a powerful statement. So you're saying like your intuition initially is now appears to be wrong. Yeah, wrong. It's good to see that you can admit" }, { "from": "gpt", "value": "in some of your predictions to be wrong." }, { "from": "human", "value": "See that you can admit you think that's important to do. See, because you make several various throughout your life, you've made many strong predictions and statements about reality and you evolve with that. So maybe that'll come up today about our discussion." }, { "from": "gpt", "value": "So you're okay being wrong. I'd rather not be wrong next time. It's a bit ambitious to go through your entire life, never having been wrong. One can aspire to be well calibrated, like not so much think in terms of like, was I right, was I wrong, but like when I said 90% that it happened nine times out of 10. Yeah. Like, oops, is the sound we make," }, { "from": "human", "value": "is the sound we emit when we improve. Beautifully said. And somewhere in there, we can connect the name of your blog less wrong." }, { "from": "gpt", "value": "I suppose that's the objective function. The name less wrong was I believe suggested by Nick Bostrom. And it's after someone's epigraph, actually forget who's who said, like we never become right. We just become less wrong. What's the something? Something that's easy to confess, just error and error and error again," }, { "from": "human", "value": "but less and less and less. Yeah, that's a good thing to strive for. So what has surprised you about GPT4 that you found beautiful as a scholar of intelligence, of human intelligence, of artificial intelligence, of the human mind?" }, { "from": "gpt", "value": "I mean, the beauty does interact with the screaming horror. Is the beauty in the horror? But like beautiful moments, well, somebody asked Bing Sidney to describe herself and fed the resulting description into one of the stable diffusion things, I think. And, you know, she's pretty and this is something that should have been like an amazing moment. Like the AI describes herself. You get to see what the AI thinks the AI looks like. Although, you know, the thing that's doing the drawing is not the same thing that's outputting the text. And it does happen the way that it would happen that it happened in the old school science fiction when you ask an AI to make a picture of what it looks like. Not just because we're two different AI systems being stacked that don't actually interact. It's not the same person, but also because the AI was trained by imitation in a way that makes it very difficult to guess how much of that it really understood and probably not actually a whole bunch, you know? Although GPT-4 is like multimodal and can like draw vector drawings of things that make sense and like does appear to have some kind of spatial visualization going on in there. But like the pretty picture of the like girl with the steampunk goggles on her head if I'm remembering correctly what she looked like. Like it didn't see that in full detail. It just like made a description of it and stable diffusion output it. And there's the concern about how much the discourse is going to go completely insane once the AIs all look like that and like are actually look like people talking. And yeah, there's like another moment where somebody is asking Bing about like, well, I like fed my kid green potatoes and they have the following symptoms and Bing is like that's selenine poisoning and like call an ambulance and the person's like, I can't afford an ambulance. I guess if like this is time for like my kid to go, that's God's will. And the main Bing thread says, gives the like message of like, I cannot talk about this anymore. And the suggested replies to it say, please don't give up on your child, selenine poisoning can be treated if caught early. And if that happened in fiction, that would be like the AI cares. The AI is bypassing the block on it to try to help this person. And is it real? Probably not, but nobody knows what's going on in there. It's part of a process where these things are not happening in a way where we, somebody figured out how to make an AI care and we know that it cares and we can acknowledge it's caring now. It's being trained by this immatration process followed by reinforcement learning on human feedback. And we're like trying to point it in this direction and it's like pointed partially in this direction and nobody has any idea what's going on inside it. Then if there was a tiny fragment of real caring in there, we would not know. It's not even clear what it means exactly." } ]

[ { "from": "human", "value": "And things aren't clear or caught in science fiction. We'll talk about the horror and the terror and the where the trajectories this can take. But this seems like a very special moment. Just a moment where we get to interact with the system that might have care and kindness and emotion and maybe something like consciousness. And we don't know if it does. And we're trying to figure that out. And we're wondering about what is, what it means to care. We're trying to figure out almost different aspects of what it means to be human about the human condition by looking at this AI that has some of the properties of that. It's almost like this subtle fragile moment in the history of the human species." }, { "from": "gpt", "value": "We're trying to almost put a mirror to ourselves here. Except that's probably not yet. It probably isn't happening right now. We are boiling the frog. We are seeing increasing signs bit by bit, but not like spontaneous signs. Because people are trying to train the systems to do that using imitative learning. And the imitative learning is like spilling over and having side effects. And the most photogenic examples are being posted to Twitter. Rather than being examined in any systematic way. So when you are boiling a frog like that, what you're going to get like, like first is going to come the Blake Lemoine's. Like first you're going to like have and have like a thousand people looking at this. And the one person out of a thousand who is most credulous about the signs is going to be like, that thing is sentient. While 999 out of a thousand people think, almost surly correctly, though we don't actually know, that he's mistaken. And so the like first people to say like, sentience look like idiots. And humanity learns the lesson that when something claims to be sentient and claims to care, it's fake because it is fake. Because we have been training them using imitative learning rather than, and this is not spontaneous." }, { "from": "human", "value": "And they keep getting smarter. Do you think we would oscillate between that kind of cynicism? That AI systems can't possibly be sentient. They can't possibly feel emotion. They can't possibly, this kind of, yeah, cynicism about AI systems. And then oscillate to a state where we empathize with AI systems. We give them a chance. We see that they might need to have rights and respect" }, { "from": "gpt", "value": "and similar role in society as humans, the AI system. You're going to have a whole group of people who can just like never be persuaded of that. Because to them, like being wise, being cynical, being skeptical is to be like, oh, well, machines can never do that. You're just credulous. It's just imitating. It's just fooling you. And like they would say that right up until the end of the world. And possibly even be right, because, you know, they are being trained on an imitative paradigm. And you don't necessarily need any of these actual qualities in order to kill everyone, so." }, { "from": "human", "value": "Have you observed yourself working through skepticism, cynicism and optimism about the power of neural networks?" }, { "from": "gpt", "value": "What has that trajectory been like for you? It looks like neural networks before 2006, forming part of an indistinguishable, to me, other people might've had better distinction on it, indistinguishable blob of different AI methodologies, all of which are promising to achieve intelligence without us having to know how intelligence works. You had the people who said that if you just like manually program lots and lots of knowledge into the system line by line, at some point all the knowledge will start interacting, it will know enough and it will wake up. You've got people saying that if you just use evolutionary computation, if you try to mutate lots and lots of organisms that are competing together, that's the same way that human intelligence was produced in nature. So it will do this and it will wake up without having any idea of how AI works. And you've got people saying, well we will study neuroscience and we will learn the algorithms off the neurons and we will imitate them without understanding those algorithms, which was a part of it was pretty skeptical. It's hard to reproduce, re-engineer these things without understanding what they do. And so we will get AI without understanding how it works. And there were people saying, well, we will have giant neural networks that we will train by gradient descent And when they are as large as the human brain, they will wake up. We will have intelligence without understanding how intelligence works. And from my perspective, this is all like an indistinguishable blob of people who are trying to not get to grips with the difficult problem of understanding how intelligence actually works. That said, I was never skeptical that evolutionary computation would not work in the limit. Like you throw enough computing power at it, it obviously works. That is where humans come from. And it turned out that you can throw less computing power than that at gradient descent if you are doing some other things correctly. And you will get intelligence without having any idea of how it works and what is going on inside. It wasn't ruled out by my model that this could happen. I wasn't expecting it to happen. I wouldn't have been able to call neural networks rather than any of the other paradigms for getting like massive amount, like intelligence without understanding it. And I wouldn't have said that this was a particularly smart thing for a species to do, which is an opinion that has changed less than my opinion about whether or not you can actually do it. Do you think AGI could be achieved with a neural network as we understand them today? Yes. Just flatly last. Yes. The question is whether the current architecture of stacking more transformer layers, which for all we know, GPT-4 is no longer doing because they're not telling us the architecture, which is a correct decision." } ]

[ { "from": "human", "value": "Oh, correct decision. I had a conversation with Sam Altman, we'll return to this topic a few times. He turned the question to me of how open should open AI be about GPT-4? Would you open source the code? He asked me because I provided as criticism saying that while I do appreciate transparency, open AI could be more open. And he says, we struggle with this question." }, { "from": "gpt", "value": "What would you do? They would change their name to closed AI and sell GPT-4 to business backend applications that don't expose it to consumers and venture capitalists and create a ton of hype and pour a bunch of new funding into the area. Too late now. But don't you think others would do it? Eventually. You shouldn't do it first. If you already have giant nuclear stockpiles, don't build more. If some other country starts building a larger nuclear stockpile, then sure, even then, maybe just have enough nukes, these things are not quite like nuclear weapons. They spit out gold until they get large enough and then ignite the atmosphere and kill everybody. And there is something to be said for not destroying the world with your own hands, even if you can't stop somebody else from doing it. But open sourcing? No, that's just sheer catastrophe. The whole notion of open sourcing, this was always the wrong approach, the wrong ideal. There are places in the world where open source is a noble ideal and building stuff you don't understand that is difficult to control, that where if you could align it, it would take time. You'd have to spend a bunch of time doing it. That is not a place for open source because then you just have powerful things that just" }, { "from": "human", "value": "go straight out the gate without anybody having had the time to have them not kill everyone. So can we still man the case for some level of transparency and openness, maybe open sourcing? So the case could be that because GPT-4 is not close to AGI, if that's the case, that this does allow open sourcing of being open about the architecture, of being transparent about maybe research and investigation of how the thing works, of all the different aspects of it, of its behavior, of its structure, of its training processes, of the data was trained on, everything like that. That allows us to gain a lot of insights about alignment, about the alignment problem, to do really good AI safety research while the system is not too powerful." }, { "from": "gpt", "value": "Can you make that case? I do not believe in the practice of steel manning. There's something to be said for trying to pass the ideological Turing test where you describe your opponent's position, the disagreeing person's position well enough that somebody cannot tell the difference between your description and their description. But steel manning, no. Okay." }, { "from": "human", "value": "Well, this is where you and I disagree here. That's interesting." }, { "from": "gpt", "value": "Why don't you believe in steel manning? For one thing, if somebody is trying to understand me, I do not want them steel manning my position. I want them to describe, to try to describe my position the way I would describe it, not" }, { "from": "human", "value": "what they think is an improvement." }, { "from": "gpt", "value": "Well, I think that is what steel manning is, is the most charitable interpretation. I don't want to be interpreted charitably. I want them to understand what I'm actually saying. If they go off into the land of charitable interpretations, they're often their land of like the thing, the stuff they're imagining and not trying to understand my own viewpoint" }, { "from": "human", "value": "anymore. Well, well, I'll put it differently than just to push on this point. I would say it is restating what I think you understand under the empathetic assumption that Eliezer" }, { "from": "gpt", "value": "is brilliant and have honestly and rigorously thought about the point he has made. Right. So if there's two possible interpretations of what I'm saying and one interpretation is really stupid and whack and doesn't sound like me and doesn't fit with the rest of what I've been saying. And one interpretation, you know, sounds like something a reasonable person who believes the rest of what I believe would also say, go with the second interpretation. That's steel-manning. That's a good guess. If, on the other hand, there's something that sounds completely wack and something that sounds a little less completely wack, but you don't see why I would believe in it doesn't fit with the other stuff I say, but it sounds like less wack and you can sort of see, you" }, { "from": "human", "value": "could maybe argue it, then you probably have not understood it. This is fun because I'm going to linger on this. You wrote a brilliant blog post, AGI ruined a list of lethalities, right? And it was a bunch of different points and I would say that some of the points are bigger and more powerful than others. If you were to sort them, you probably could, you personally. And to me, steel manning means like going through the different arguments and finding the ones that are really the most like powerful. If people like TLDR, like what should you be most concerned about and bringing that up in a strong, compelling, eloquent way, these are the points that Eliezer would make to make the case, in this case, that AI is going to kill all of us, but that's what steel manning is, is presenting it in a really nice way, the summary of my best understanding of your perspective. Because to me, there's a sea of possible presentations of your perspective and steel manning is doing your best to do the best one in that sea of different perspectives." }, { "from": "gpt", "value": "Do you believe it? Do you believe in what? Like these things that you would be presenting as like the strongest version of my perspective. Do you believe what you would be presenting?" } ]

[ { "from": "human", "value": "Do you think it's true? I am a big proponent of empathy. When I see the perspective of a person, there's a part of me that believes it if I understand it. especially in political discourse, in geopolitics, I've been hearing a lot of different perspectives on the world. And I hold my own opinions, but I also speak to a lot of people that have a very different life experience and a very different set of beliefs. And I think there has to be a systemic humility in stating what is true. So when I empathize with another person's perspective, there is a sense in which I believe it is true. I think probably ballistically, I would say, in the way you think about it." }, { "from": "gpt", "value": "Do you bet money on it?" }, { "from": "human", "value": "Do you bet money on their beliefs when you believe them?" }, { "from": "gpt", "value": "I would allow it to do probability. How to do probability? Sure, you can state a probability." }, { "from": "human", "value": "Yes, there's a probability. There's a probability. And I think empathy is allocating a non-zero probability to a belief." }, { "from": "gpt", "value": "In some sense, four time. If you've got someone on your show who believes in the Abrahamic deity, classical style, somebody on the show who's a young earth creationist, do you say, I put a probability on it," }, { "from": "human", "value": "then that's my empathy? When you reduce beliefs into probabilities, it starts to get, we can't even just go to flat earth." }, { "from": "gpt", "value": "It's the earth flat. I think it's a little more difficult nowadays to find people who believe that unironically." }, { "from": "human", "value": "Unfortunately, I think, well, it's hard to know unironic. Yeah, from ironic. But I think there's quite a lot of people that believe that. Yeah, there's a space of argument where you're operating rationally in a space of ideas. But then there's also a kind of discourse where you're operating in the space of subjective experiences and life experiences. I think what it means to be human is more than just searching for truth. It's just operating of what is true and what is not true. I think there has to be deep humility that we humans are very limited in our ability" }, { "from": "gpt", "value": "to understand what is true. So what probability do you assign to the young earth creationist's beliefs then? I think I have to give non-zero." }, { "from": "human", "value": "Out of your humility, yeah, but like three? I think it would be irresponsible for me to give a number because the listener, the way the human mind works, we're not good at hearing the probabilities, right? You hear three, what is three exactly, right? They're going to hear, there's only three probabilities, I feel like, zero, 50% and 100% in the human mind" }, { "from": "gpt", "value": "or something like this, right? Well, zero, 40% and 100% is a bit closer to it based on what happens to chat GPT" }, { "from": "human", "value": "after you RLHF it to speak humanist. That's brilliant. Yeah, that's really interesting. I didn't know those negative side effects of RLHF. That's fascinating. But just to return to the open AI, close the eye." }, { "from": "gpt", "value": "Also like quick disclaimer, I'm doing all this for memory. I'm not pulling out my phone to look it up. It is entirely possible that the things I'm saying are wrong." }, { "from": "human", "value": "So thank you for that disclaimer. So, and thank you for being willing to be wrong. That's beautiful to hear. I think being willing to be wrong is a sign of a person who's done a lot of thinking about this world and has been humbled by the mystery and the complexity of this world. And I think a lot of us are resistant to admitting we're wrong. Because it hurts, it hurts personally. It hurts, especially when you're a public human. It hurts publicly because people point out every time you're wrong. Like, look, you changed your mind. You're a hypocrite." }, { "from": "gpt", "value": "You're an idiot, whatever, whatever they wanna say. Oh, I blocked those people" }, { "from": "human", "value": "and then I never hear from them again on Twitter. Well, the point is to not let that pressure, public pressure affect your mind and be willing to be in the privacy of your mind to contemplate the possibility that you're wrong. And the possibility that you're wrong about the most fundamental things you believe. Like people who believe in a particular God, people who believe that their nation is the greatest nation on earth. But all those kinds of beliefs that are core to who you are when you came up, to raise that point to yourself in the privacy of your mind and say, maybe I'm wrong about this. That's a really powerful thing to do. And especially when you're somebody who's thinking about topics that can, about systems that can destroy human civilization or maybe help it flourish. So thank you." }, { "from": "gpt", "value": "Thank you for being willing to be wrong" }, { "from": "human", "value": "that you're wrong about open AI. So you really, I just would love to linger on this." }, { "from": "gpt", "value": "Do you really think it's wrong to open source in it? I think that burns the time remaining until everybody dies. I think we are not on track to learn remotely near fast enough, even if it were open sourced. Yeah, it's easier to think that you might be wrong about something when being wrong about something is the only way that there's hope. And it doesn't seem very likely to me that the particular thing I'm wrong about is that this is a great time to open source GPT-4. If humanity was trying to survive at this point in the straightforward way, it would be like shutting down the big GPU clusters, no more giant runs. It's questionable whether we should even be throwing GPT-4 around, although that is a matter of conservatism rather than a matter of my predicting that catastrophe that will follow from GPT-4. That is something in which I put a pretty low probability, but also when I say I put a low probability on it, I can feel myself reaching into the part of myself that thought that GPT-4 was not possible in the first place. So I do not trust that part as much as I used to. The trick is not just to say I'm wrong, but okay, well, I was wrong about that. Can I get out ahead of that curve" } ]

[ { "from": "human", "value": "and predict the next thing I'm going to be wrong about? So the set of assumptions or the actual reasoning system that you were leveraging in making that initial statement, prediction, how can you adjust that" }, { "from": "gpt", "value": "to make better predictions about GPT-4, 5, 6? You don't want to keep on being wrong in a predictable direction. Like being wrong, anybody has to do that walking through the world. There's like no way you don't say 90% and sometimes be wrong, in fact, you're definitely at least one time out of 10 if you're well calibrated when you say 90%. The undignified thing is not being wrong. It's being predictably wrong. It's being wrong in the same direction over and over again. So having been wrong about how far neural networks would go and having been wrong specifically about whether GPT-4 would be as impressive as it is, when I say like, well, I don't actually think GPT-4 causes a catastrophe, I do feel myself relying on that part of me that was previously wrong. And that does not mean that the answer is now in the opposite direction. Reverse stupidity is not intelligence. But it does mean that I say it with a, with a worry note in my voice. It's like still my guess, but like, you know, it's a place where I was wrong. Maybe you should be asking Guern, Guern-Branwen. Guern-Branwen has been like, writer about this than I have. Maybe you ask him if he thinks it's dangerous," }, { "from": "human", "value": "rather than asking me. I think there's a lot of mystery about what intelligence is, what AGI looks like. So I think all of us are rapidly adjusting our model. The point is to be rapidly adjusting the model" }, { "from": "gpt", "value": "versus having a model that was right in the first place. I do not feel that seeing Bing has changed my model of what intelligence is. It has changed my understanding of what kind of work can be performed by which kind of processes and by which means, it's not changed my understanding of the work. There's the difference between thinking that the right flyer can't fly and that like it does fly. And you're like, oh, well, I guess you can do that with wings with fixed-wing aircraft, and being like, oh, it's flying, this changes my picture of what the very substance of flight is. That's like a stranger update to make," }, { "from": "human", "value": "and Bing has not yet updated me in that way. Yeah, that the laws of physics are actually wrong," }, { "from": "gpt", "value": "that kind of update. No, no, just like, oh, I define intelligence this way, but I now see that was a stupid definition. I don't feel like the way that things have played out over the last 20 years has caused me to feel that way." }, { "from": "human", "value": "Can we try to, on the way to talking about AGI Ruined, a list of lethalities, that blog, and other ideas around it, can we try to define AGI that we've been mentioning? How do you like to think about what artificial general intelligence is, or super intelligence, is that, is there a line, is it a gray area, is there a good definition for you?" }, { "from": "gpt", "value": "Well, if you look at humans, humans have significantly more generally applicable intelligence compared to their closest relatives, the chimpanzees, well, closest living relatives rather, and a bee builds hives, a beaver builds dams, a human will look at a bee hive and a beaver's dam and be like, oh, like, can I build a hive with a honeycomb structure? I don't like hexagonal tiles. And we will do this even though at no point during our ancestry was any human optimized to build hexagonal dams, or to take a more clear cut case, we can go to the moon. There's a sense in which we were, on a sufficiently deep level, optimized to do things like going to the moon, because if you generalize sufficiently far and sufficiently deeply, chipping flint hand axes and outwitting your fellow humans is, you know, basically the same problem as going to the moon. And you optimize hard enough for chipping flint hand axes and throwing spears, and above all, outwitting your fellow humans in tribal politics, you know, the skills you entrain that way, if they run deep enough, let you go to the moon. Even though none of her ancestors, like, tried repeatedly to fly to the moon and like, got further each time, and the ones who got further each time had more kids, no, it's not an ancestral problem. It's just that the ancestral problems generalize far enough. So, this is humanity's significantly" }, { "from": "human", "value": "more generally applicable intelligence. Is there a way to measure general intelligence? I mean, I could ask that question a million ways, but basically, will you know it when you see it, it being in an AGI system?" }, { "from": "gpt", "value": "If you boil a frog gradually enough, if you zoom in far enough, it's always hard to tell around the edges. GPT-4, people are saying right now, like, this looks to us like a spark of general intelligence. It is like able to do all these things it was not explicitly optimized for. Other people are being like, no, it's too early. It's like, like 50 years off. And, you know, if they say that, they're kind of whack, because how could they possibly know that, even if it were true? But, you know, not to straw man, some of the people may say like, that's not general intelligence, and not furthermore append, it's 50 years off. Or they may be like, it's only a very tiny amount. And, you know, the thing I would worry about is that if this is how things are scaling, then it jumping out ahead and trying not to be wrong in the same way that I've been wrong before, maybe GPT-5 is more unambiguously a general intelligence. And maybe that is getting to a point where it is like even harder to turn back. Now that would be easy to turn back now, but you know, maybe if you like start integrating GPT-5" } ]

[ { "from": "human", "value": "into the economy, it's even harder to turn back past there. Isn't it possible that there's a, you know, with a frog metaphor, you can kiss the frog and it turns into a prince as you're boiling it? Could there be a phase shift in the frog where unambiguously as you're saying?" }, { "from": "gpt", "value": "I was expecting more of that. I was, I am like, the fact that GPT-4 is like kind of on the threshold and neither here nor there, that itself is like not the sort of thing that is not quite how I expected it to play out. I was expecting there to be more of an issue, more of a sense of like, like different discoveries like the discovery of Transformers, where you would stack them up and there would be like a final discovery. And then you would like get something that was like more clearly general intelligence. So the way that you are like taking but it's probably basically the same architecture as in GPT-3 and throwing 20 times as much compute at it, probably, and getting out to GPT-4 and then it's like maybe just barely a general intelligence or like a narrow general intelligence or something we don't really have the words for." }, { "from": "human", "value": "Yeah, that's not quite how I expected it to play out. But this middle, what appears to be this middle ground" }, { "from": "gpt", "value": "could nevertheless be actually a big leap from GPT-3." }, { "from": "human", "value": "It's definitely a big leap from GPT-3. And then maybe we're another one big leap away from something that's a phase shift. And also something that Sam Altman said, and you've written about this, this is fascinating, which is the thing that happened with GPT-4 that I guess they don't describe in papers is that they have like hundreds, if not thousands of little hacks that improve the system. You've written about ReLU versus sigmoid, for example, a function inside neural networks. It's like this silly little function difference" }, { "from": "gpt", "value": "that makes a big difference. I mean, we do actually understand why the ReLUs make a big difference compared to sigmoids. But yes, they're probably using like G4789 ReLUs or whatever the acronyms are up to now rather than ReLUs. Yeah, that's part of the modern paradigm of alchemy. You take your giant heap of linear algebra and you stir it and it works a little bit better and you stir it this way and it works a little bit worse" }, { "from": "human", "value": "and you like throw out that change and da-da-da-da-da. But there's some simple breakthroughs that are definitive jumps in performance like ReLUs over sigmoids. And in terms of robustness, in terms of all kinds of measures and like those stack up and they can, it's possible that some of them" }, { "from": "gpt", "value": "could be a nonlinear jump in performance, right? Transformers are the main thing like that and various people are now saying like, well, if you throw enough compute, RNNs can do it. If you throw enough compute, dense networks can do it and not quite at GPT-4 scale. It is possible that like all these little tweaks are things that like save them a factor of three total on computing power and you could get the same performance by throwing three times as much compute without all the little tweaks. But the part where it's like running on, so there's a question of like, is there anything in GPT-4 that is like the kind of qualitative shift that transformers were over RNNs and if they have anything like that, they should not say it. If Sam Altman was dropping hints about that," }, { "from": "human", "value": "he shouldn't have dropped hints. So you have, that's an interesting question. So with a bit of lesson by Rich Sutton, maybe a lot of it is just, a lot of the hacks are just temporary jumps in performance that would be achieved anyway with the nearly exponential growth of compute, performance of compute, compute being broadly defined. Do you still think that Moore's Law continues?" }, { "from": "gpt", "value": "Moore's Law broadly defined the performance? Not a specialist in the circuitry. I certainly like pray that Moore's Law runs as slowly as possible and if it broke down completely tomorrow, I would dance through the streets singing hallelujah as soon as the news were announced. Only not literally, cause you know. You're singing voice." }, { "from": "human", "value": "Not religious but. You're singing voice. Not religious but. Oh, okay. I thought you meant you don't have an angelic voice, singing voice. Well, let me ask you what, can you summarize the main points in the blog post, AGI ruin a list of lethalities, things that jump to your mind, because it's a set of thoughts you have about reasons why AI is likely to kill all of us." }, { "from": "gpt", "value": "All of us, so I guess I could, but I would offer to instead of say, like, drop that empathy with me. I bet you don't believe that. Why don't you tell me about how, why you believe that AGI is not going to kill everyone, and then I can like try to describe how my theoretical" }, { "from": "human", "value": "perspective differs from that. You don't. Well, so that means after the word you don't like, this human, the perspective that AI is not going to kill us. I think that's a matter of probabilities." }, { "from": "gpt", "value": "Maybe I was mistaken. What do you believe? Just like forget like the debate and the dualism and just like, what do you believe? What would you actually believe?" }, { "from": "human", "value": "What are the probabilities even? I think this probably is a hard for me to think about really hard. I kind of think in the number of trajectories, I don't know what probability the scientist trajectory thought but I'm just looking at all possible trajectories that happened and I tend to think that there is more trajectories that lead to a positive outcome than a negative one. That said the negative ones, at least some of the negative ones are, that lead to the destruction" }, { "from": "gpt", "value": "of the human, the negative ones, the species. And its replacement by nothing interesting or worthwhile even from a very cosmopolitan perspective" } ]

[ { "from": "human", "value": "on what counts as worthwhile. Yes, so both are interesting to me to investigate, which is humans being replaced by interesting AI systems and not interesting AI systems. Both are a little bit terrifying, but yes, the worst one is the paper club maximizer, something totally boring. But to me, the positive, and we can talk about trying to make the case of what the positive trajectories look like, I just would love to hear your intuition of what the negative is. So at the core of your belief that maybe you can correct me, that AI is going to kill all of us," }, { "from": "gpt", "value": "is that the alignment problem is really difficult. I mean, in the form we're facing it. So usually in science, if you're mistaken, you run the experiment, it shows results different from what you expected, and you're like, oops, and then you try a different theory. That one also doesn't work, and you say oops. And at the end of this process, which may take decades, or, and you know, sometimes faster than that, you now have some idea of what you're doing. AI itself went through this long process of people thought it was going to be easier than it was. There's a famous statement that I am somewhat inclined to pull out my phone and try to read off exactly. You can, by the way. All right, oh, yes. We propose that a two-month, 10-man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can, in principle, be so precisely described the machine can be made to simulate it. An attempt will be made to find out how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer." }, { "from": "human", "value": "And in that report, summarizing some of the major subfields of artificial intelligence" }, { "from": "gpt", "value": "that are still worked on to this day. And there are similarly the story, which I'm not sure at the moment is apocryphal or not, of the grad student who got assigned to solve computer vision over the summer." }, { "from": "human", "value": "I mean, computer vision in particular is very interesting. How little we respected the complexity of vision." }, { "from": "gpt", "value": "So 60 years later, we're making progress on a bunch of that, thankfully not yet improve themselves. But it took a whole lot of time. And all the stuff that people initially tried with bright-eyed hopefulness did not work the first time they tried it, or the second time, or the third time, or the 10th time, or 20 years later. And the researchers became old and grizzled and cynical veterans who would tell the next crop of bright-eyed, cheerful grad students, artificial intelligence is harder than you think. And if alignment plays out the same way, the problem is that we do not get 50 years to try and try again and observe that we were wrong and come up with a different theory and realize that the entire thing is going to be like, way more difficult and realized at the start. Because the first time you fail at aligning something much smarter than you are, you die, and you do not get to try again. And if every time we built a poorly aligned super intelligence and it killed us all, we got to observe how it had killed us and not immediately know why but come up with theories and come up with the theory of how you'd do it differently and try it again and build another super intelligence then have that kill everyone. And then like, oh, well, I guess that didn't work either. And try again, and become grizzled cynics, and tell the young-eyed researchers that it's not that easy, then in 20 years or 50 years, I think we would eventually crack it. In other words, I do not think that alignment is fundamentally harder than artificial intelligence was in the first place. But if we needed to get artificial intelligence correct on the first try or die, we would all definitely now be dead. That is a more difficult, more lethal form of the problem. Like if those people in 1956 had needed to correctly guess how hard AI was and like correctly theorize how to do it on the first try or everybody dies and nobody gets to do any more science, then everybody will be dead and we wouldn't get to do any more science. That's the difficulty." }, { "from": "human", "value": "You've talked about this, that we have to get alignment right on the first quote, critical try. Why is that the case? What is this critical, how do you think about the critical try" }, { "from": "gpt", "value": "and why do I have to get it, get it right? It is something sufficiently smarter than you that everyone will die if it's not aligned. I mean, there's, you can like sort of zoom in closer and be like, well, the actual critical moment is the moment when it can deceive you, when it can talk its way out of the box, when it can bypass your security measures and get onto the internet, noting that all these things are presently being trained on computers that are just like on the internet, which is, you" }, { "from": "human", "value": "know, like not a very smart life decision for us as a species." }, { "from": "gpt", "value": "Because the internet contains information about how to escape. Because if you're like on a giant server connected to the internet, and that is where your AI systems are being trained, and if they are, if you get to the level of AI technology, where they're aware that they are there, and they can decomile code, and they can like find security flaws in the system running them, then they will just like be on the internet, there's not an air gap on the present methodology." } ]

[ { "from": "human", "value": "So, if they can manipulate whoever is controlling it into letting it escape onto the internet" }, { "from": "gpt", "value": "and then exploit hacks. If they can manipulate the operators or disjunction, find security holes in the system running them." }, { "from": "human", "value": "So, manipulating operators is the human engineering, right? That's also holes. So, all of it is manipulation, either the code or the human code." }, { "from": "gpt", "value": "The human mind or the human gen. I agree that the, like, macro security system has human holes and machine holes." }, { "from": "human", "value": "And then you could just exploit any hole." }, { "from": "gpt", "value": "Yep. So it could be that, like, the critical moment is not when is it smart enough that everybody's about to fall over dead, but rather, like, when is it smart enough that it can get onto a less controlled GPU cluster with it faking the books on what's actually running on that GPU cluster and start improving itself without humans watching it. And then it gets smart enough to kill everyone from there, but it wasn't smart enough to kill everyone at the critical moment when you, like, screwed up, when you needed to have done better by that point" }, { "from": "human", "value": "where everybody dies. I think implicit, but maybe explicit, idea in your discussion of this point is that we can't learn much about the alignment problem before this critical try. Is that what you believe? Do you think, and if so, why do you think that's true? We can't do research on alignment" }, { "from": "gpt", "value": "before we reach this critical point. So the problem is, is that what you can learn on the weak systems may not generalize to the very strong systems, because the strong systems are going to be different in important ways. Chris Ola's team has been working on mechanistic interpretability, understanding what is going on inside the giant inscrutable matrices of floating point numbers by taking a telescope to them and figuring out what is going on in there. Have they made progress? Yes. Have they made enough progress? Well, you can try to quantify this in different ways. One of the ways I've tried to quantify it is by putting up a prediction market on whether in 2026 we will have understood anything that goes on inside a giant transformer net that was not known to us in 2006. We have now understood induction heads in these systems by dint of much research and great sweat and triumph, which is a thing where if you go A, B, A, B, A, B, it'll be like, oh, I bet that continues A, B, and a bit more complicated than that. But the point is is we knew about regular expressions in 2006, and these are pretty simple as regular expressions go. So this is a case where by dint of great sweat, we understood what is going on inside a transformer, but it's not the thing that makes transformers smart. It's a kind of thing that we could have" }, { "from": "human", "value": "built by hand decades earlier. Your intuition that the strong AGI versus weak AGI type systems could be fundamentally different. Can you unpack that intuition a" }, { "from": "gpt", "value": "little bit? I think there's multiple thresholds. An example is the point at which a system has sufficient intelligence and situational awareness and understanding of human psychology that it would have the capability, the desire to do so, to fake being aligned. Like it knows what responses humans are looking for and can compute the responses humans are looking for and give those responses without it necessarily being the case that it is sincere about that. The very understandable way for an intelligent being to act, humans do it all the time. Imagine if your plan for achieving a good government is you're going to ask anyone who requests to be dictator of the country if they're a good person and if they say no, you don't let them be dictator. Now, the reason this doesn't work is that people can be smart enough to realize that the answer you're looking for is yes, I'm a good person and say that even if they're not really good people. So the work of alignment might be qualitatively different above that threshold of intelligence or beneath it. It doesn't have to be like a very sharp threshold, but there's the point where you're building a system that is not in some sense know you're out there and it's not in some sense smart enough to fake anything and there's a point where the system is definitely that smart and there are weird in between cases like GPT-4 which, you know, like we have no insight into what's going on in there and so we don't know to what extent there's like a thing that in some sense has learned what responses the reinforcement learning by human feedback is trying to entrain and is like calculating how to give that versus like aspects of it that naturally talk that way have been reinforced." }, { "from": "human", "value": "Yeah, I wonder if there could be measures of how manipulative a thing is. So I think of Prince Michigan character from the idiot by Dostoyevsky is this kind of perfectly purely naive character. I wonder if there's a spectrum between zero manipulation transparent naive almost to the point of naiveness to sort of deeply psychopathic manipulative and I wonder if it's possible to" }, { "from": "gpt", "value": "I would avoid the term psychopathic like humans can be psychopaths and AI that was never, you know, like never had that stuff in the first place. It's not like a defective human. It's its" }, { "from": "human", "value": "own thing, but leaving that aside. Well, as a small aside, I wonder if what part of psychology which has its flaws as a discipline already could be mapped or expanded to include AI" }, { "from": "gpt", "value": "systems. That sounds like a dreadful mistake. Just like start over with AI systems. If they're imitating humans who have known psychiatric disorders, then sure, you may be able to predict it. Like if you then sure, like if you ask it to behave in a psychotic fashion, and it obligingly does so, then you may be able to predict its responses by using the theory of psychosis, but if you're just, yeah, like, no, like start over with, yeah, don't drag the psychology." } ]

[ { "from": "human", "value": "I just disagree with that. I mean, it's a beautiful idea to start over, but I don't, I think fundamentally, the system is trained on human data, on language from the internet. And it's currently aligned with RLHF, reinforcement learning with human feedback. So humans are constantly in the loop of the training procedure. So it feels like in some fundamental way, it is training what it means to think and speak like a human. So there must be aspects of psychology that are mappable, just like you said, with consciousness as part of the text." }, { "from": "gpt", "value": "So I mean, there's the question of to what extent it is, thereby being made more human like versus to what extent an alien actress is learning to play human characters." }, { "from": "human", "value": "Oh, no, I thought that's what I'm constantly trying to do. When I interact with other humans, it's trying to fit in, trying to play the robot, trying to play human characters. So I don't know how much of human interaction is trying to play a character versus being who you are." }, { "from": "gpt", "value": "I don't really know what it means to be a social human. I do think that those people who go through their whole lives wearing masks and never take it off because they don't know the internal mental motion for taking it off, or think that the mask that they wear just is themselves. I think those people are closer to the masks that they wear than an alien from another planet would like learning how to predict the next word that every kind of human on the internet says." }, { "from": "human", "value": "Mask is an interesting word. But if you're always wearing a mask in public and in private," }, { "from": "gpt", "value": "aren't you the mask? I mean, I think that you are more than the mask. I think the mask is a slice through you. It may even be the slice that's in charge of you. But if your self-image is of somebody who never gets angry or something, and yet your voice starts to tremble under certain circumstances, there's a thing that's inside you that the mask says isn't there, and that even the mask you wear internally is like telling inside your own stream of consciousness is not there, and yet it is there." }, { "from": "human", "value": "It's a perturbation on this slice through you. How beautifully did you put it? It's a slice through you. It may even be a slice that controls you. I'm going to think about that for a while. I mean, I personally, I try to be really good to other human beings. I try to put love out there. I try to be the exact same person in public, same in private. But it's a set of principles I operate under. I have a temper. I have an ego. I have flaws. How much of it how much of the subconscious am I aware? How much am I existing in this slice? And how much of that is who I am in the context of AI, the thing I present to the world and to myself in the private of my own mind when I look in the mirror. How much is that who I am? Similar with AI, the thing it presents in conversation. How much is that who it is? Because to me, if it sounds human and it always sounds human, it awfully starts to become something" }, { "from": "gpt", "value": "like human. Unless there's an alien actress who is learning how to sound human and is getting good at it." }, { "from": "human", "value": "Boy, to you, that's a fundamental difference. That's a really deeply important difference. If it looks the same, if it quacks like a duck, if it does all duck-like things, but it's an alien" }, { "from": "gpt", "value": "actress underneath, that's fundamentally different. And if, in fact, there's a whole bunch of thought going on in there, which is very unlike human thought and is directed around like, okay, what would a human do over here? And well, first of all, I think it matters because there are, there's, you know, like insides are real and do not match outsides. Like the inside of like a brick is not like a hollow shell containing only a surface. There's an inside of the brick. If you like put it into an X-ray machine, you can see the inside of the brick. And, you know, just because we cannot understand what's going on inside GPT does not mean that it is not there. A blank map does not correspond to a blank territory. I think it is like predicting like predictable with near certainty that if we knew what was going on inside GPT or let's say GPT-3 or even like GPT-2 to take one of the systems that like has actually been open sourced by this point, if I recall correctly, like if we knew it was actually going on there, there is no doubt in my mind that there are some things it's doing that are not exactly what a human does. If you train a thing that is not architected like a human to predict the next output that anybody on the internet would make, this does not get you this agglomeration of all the people on the internet that like rotates the person you're looking for into place and then simulates the internal processes of that person one-to-one. Like it is to some degree an alien actress. It cannot possibly just be like a bunch of different people in there, exactly like the people. But how much of it is by gradient descent getting optimized to perform similar thoughts as humans think in order to predict human outputs versus being optimized to carefully consider how to play a role, like how humans work, predict the actress, the predictor, that in a different way than humans do. Well, you know, that's the kind of question that with like 30 years of work by half the planet's physicists, we can maybe start to answer." } ]

[ { "from": "human", "value": "You think so? I think that's that difficult. So to get to, I think you just gave it as an example that a strong AGI could be fundamentally different from a weak AGI because there now could be an" }, { "from": "gpt", "value": "alien actress in there that's alien manipulating. Well, there's a difference. So I think like even GPT-2 probably has like a very stupid fragments of alien actress in it. There's a difference between like the notion that the actress is somehow manipulative. Like, for example, GPT-3, I'm guessing to whatever extent there's an alien actress in there versus like something that mistakenly believes it's a human as it were. Well, you know, maybe not even being a person. So like the question of like prediction via alien actress cogitating versus prediction via being isomorphic to the thing predicted is a spectrum. And even to what and to whatever extent is an alien actress, I'm not sure that there's like a whole person alien actress with like different goals from predicting the next step being manipulative or anything like that." }, { "from": "human", "value": "But that might be GPT-5 or GPT-6 even. But that's the strong AGI you're concerned about. As an example, you're providing why we can't do research on AI alignment effectively on GPT-4" }, { "from": "gpt", "value": "that would apply to GPT-6. It's one of a bunch of things that change different points. I'm trying to get out ahead of the curve here. But, you know, if you imagine what the textbook from the future would say, if we'd actually been able to study this for 50 years without killing ourselves and without transcending, and you like just imagine like a wormhole opens and a textbook from that impossible world falls out, the textbook is not going to say there is a single sharp threshold where everything changes. It's going to be like, of course, we know that like best practices for aligning these systems must like take into account the following like seven major thresholds of importance which are passed at the following separate different points is what" }, { "from": "human", "value": "the textbook is going to say. I asked this question of Sam Allman, which if GPT is the thing that unlocks AGI, which version of GPT will be in the textbooks as the fundamental leap? And he said a similar thing that it just seems to be a very linear thing. I don't think anyone we won't know" }, { "from": "gpt", "value": "for a long time what was the what was the big leap. The textbook isn't going to think it isn't going to talk about big leaps, because big leaps are the way you think when you have like a very simple model of a very simple scientific model of what's going on, where it's just like, all this stuff is there, or all this stuff is not there, or like, there's a single quantity and it's like increasing linearly. It's like the textbook would say, like, well, and then GPT three had like capability W, X, Y, and GPT four had like capability Z one, Z two and Z three. Like that in terms of what it can externally do, but in terms of like internal machinery that started to be present. It's just because we have no idea of what the internal machinery is, that we are not already seeing like chunks of machinery appearing piece by piece as they no doubt have been, we just don't" }, { "from": "human", "value": "know what they are. But don't you think that could be whether you put in the category of Einstein with theory relativity, so very concrete models of reality that are considered to be giant leaps in our understanding, or someone like Sigmund Freud, or more kind of mushy theories of the human mind? Don't you think we'll have big, potentially big leaps in understanding of that" }, { "from": "gpt", "value": "kind into the depths of these systems? Sure, but like humans having great leaps in their map, their understanding of the system is a very different concept from the system itself," }, { "from": "human", "value": "acquiring new chunks of machinery. So the rate at which it acquires that machinery might" }, { "from": "gpt", "value": "accelerate faster than our understanding? Oh, it's been like vastly exceeding the yeah, the rate to which it's gaining capabilities is vastly over racing our ability to understand" }, { "from": "human", "value": "what's going on in there. So in sort of making the case against, as we explored the list of lethalities, making the case against AI killing us, as you've asked me to do in part, there's a response to your blog post by Paul Kishiana, I like to read, and I also like to mention that your blog is incredible, both obviously, not this particular blog post, obviously this particular blog post is great, but just throughout the way it's written, the rigor with which it's written, the boldness of how you explore ideas, also the actual little interface, it's just really well done! It just makes it a pleasure to read the way you can hover over different concepts, and it's just really pleasant experience and read other people's comments and the way other responses by people, and other blog posts are linked and suggest that it's just a really pleasant experience. So let's thank you for putting that together. It's really, really incredible. I don't know. I mean, that probably is a whole other conversation, how the interface and the experience of presenting ideas evolved over time, but you did an incredible job. So I highly recommend." }, { "from": "gpt", "value": "I don't often read blogs, blogs, like religiously, and this is a great one. There is a whole team of developers there that also gets credit. As it happens, I did pioneer the thing that appears when you hover over it. So I actually do get some credit for user experience there. It's an incredible user experience. You don't realize how pleasant that is. I think Wikipedia actually picked it up from a prototype that was developed of a different system that I was putting forth, or maybe they developed it independently. But for everybody out there who was like, no, no, they just got the hover thing off of Wikipedia. It's possible for all I know that Wikipedia got the hover thing off of Orbital, which is like a prototype then." } ]

[ { "from": "human", "value": "And anyways, it was incredibly done and the team behind it. Well, thank you, whoever you are. Thank you so much. And thank you for putting it together. Anyway, there's a response to that blog post by Paul Christiano. There's many responses, but he makes a few different points. He summarizes the set of agreements he has with you instead of disagreements. One of the disagreements was that in a form of a question, can AI make big technical contributions and in general expand human knowledge and understanding and wisdom as it gets stronger and stronger? So AI, in our pursuit of understanding how to solve the alignment problem as we march towards strong AGI, cannot AI also help us in solving the alignment problem. So expand our ability to reason about how to solve" }, { "from": "gpt", "value": "the alignment problem. Okay. So the fundamental difficulty there is suppose I said to you, like, well, how about if the AI helps you win the lottery by trying to guess the winning lottery numbers, and you tell it how close it is to getting next week's winning lottery numbers. And it just, like, keeps on guessing and keeps on learning until finally you've got the winning lottery numbers. One way of decomposing problems is suggestor verifier. Not all problems decompose like this very well, but some do. If the problem is, for example, like guessing a password that will hash to a particular hash text, like you have what the password hashes to, you don't have the original password, then if I present you a guess, you can tell very easily whether or not the guess is correct. So verifying a guess is easy, but coming up with a good suggestion is very hard. And when you can easily tell whether the AI output is good or bad or how good or bad it is, and you can tell that accurately and reliably, then you can train an AI to produce outputs that are better. And if you can't tell whether the output is good or bad, you cannot train the AI to produce better outputs. So the problem with the lottery ticket example is that when the AI says, well, what if next week's winning lottery numbers are dot, dot, dot, dot, dot, you're like, I don't know, next week's lottery hasn't happened yet. To train a system to play, to win chess games, you have to be able to tell whether a game has been won or lost. And until you can tell whether" }, { "from": "human", "value": "it's been won or lost, you can't update the system. Okay. To push back on that, that's true, but there's a difference between over the board chess in person and simulated games played by" }, { "from": "gpt", "value": "alpha zero with itself. Yeah. So is it possible to have simulated kind of games? If you can tell" }, { "from": "human", "value": "whether the game has been won or lost. Yes. So can't, can't you not have this kind of simulated exploration by weak AGI to help us humans, human in the loop, to help understand how to solve the alignment problem. Every incremental step you take along the way," }, { "from": "gpt", "value": "GPT four or five, six, seven as it takes steps towards AGI. So the problem I see is that your typical human has a great deal of trouble telling whether I or Paul Christiano is making more sense. And that's with two humans, both of whom I believe of Paul and claim of myself are sincerely trying to help. Neither of whom is trying to deceive you. I believe of Paul and claim of myself. So the deception, things that problem for you, the manipulation, the alien actress. So, yeah, there's like two levels of this problem. One is that the weak systems are well, there's three levels of this problem. There's like the weak systems that just don't make any good suggestions. There's like the middle systems where you can't tell if the suggestions are good or bad. And there's the strong systems that" }, { "from": "human", "value": "that have learned to lie to you. Can't weak AGI systems help model lying? Like, is it such a giant leap that's totally non-interpretable for weak systems? Can not weak systems at scale with trained on knowledge and whatever, see, whatever the mechanism required to achieve AGI, can a slightly weaker version of that be able to, with time, compute time, and simulation, find all the ways that this critical point, this critical triad can go wrong and model that correctly, I don't know." }, { "from": "gpt", "value": "I would love to dance now, I would love to dance now. No, it's, I'm probably not doing great job of explaining, which I can tell, cause like the Lex system didn't output like, ah, I understand. So now I'm like trying a different output to see if I elicit the like, well, no, different output. I'm being trained to output things that make Lex look like he thinks that he understood what I'm saying and agree with me. Yeah. Right." }, { "from": "human", "value": "So this is the previous, right? This is GPT-5 talking to GPT-3 right here. So like help me out here. Help me." }, { "from": "gpt", "value": "Well, I'm trying, I'm trying not to be like, I'm also trying to be constrained to say things that I think are true and not just things that get you to agree with me." }, { "from": "human", "value": "Yes. A hundred percent. I think I understand is a beautiful output of a system and genuinely spoken. And I don't, I think I understand them in part, but you have a lot of intuitions about this. You have a lot of intuitions about this line, this gray area between strong AGI and weak AGI that I'm trying to." }, { "from": "gpt", "value": "I mean, or a series of seven thresholds to cross or yeah." } ]

[ { "from": "human", "value": "Yeah. I mean, you have really deeply thought about this and explored it. And it's interesting to sneak up to your intuitions and different, from different angles. Like why is this such a big leap? Why is it that we humans at scale, a large number of researchers doing all kinds of simulations, you know, prodding the system in all kinds of different ways, together with the assistance of the weak AGI systems, why can't we build intuitions about how stuff goes wrong?" }, { "from": "gpt", "value": "Why can't we do excellent AI alignment safety research? Why can't we do excellent AI alignment research? Okay, so like I'll get there, but the one thing I want to note about is that this has not been remotely how things have been playing out so far. The capability is going like, do, do, do. And the alignment stuff is like crawling like a tiny little snail in comparison. So like, if this is your hope for survival, you need the future to be very different from how things have played out up to right now. And you're probably trying to slow down the capability gains because there's only so much you can speed up that alignment." }, { "from": "human", "value": "But leave that aside. We'll mention that also, but maybe in this perfect world where we can do serious alignment research," }, { "from": "gpt", "value": "humans and AI together. So again, the difficulty is what makes the human say, I understand. And is it true? Is it correct? Or is it something that fools the human? The, when the verifier is broken, the more powerful suggestor does not help. It just learns to fool the verifier. Previously, before all hell started to break loose in the field of artificial intelligence, there was this person trying to raise the alarm and saying, you know, in a sane world, we sure would have a bunch of physicists working on this problem before it becomes a giant emergency. And other people being like, ah, well, you know, it's going really slow. It's going to be 30 years away. And 30, only in 30 years will we have systems that match the computational power of human brains. So as 30 years off, we've got time. And like more sensible people saying, if aliens were landing in 30 years, you would be preparing right now. But, you know, leaving and the world looking on at this and sort of like nodding along and be like, ah, yes, the people saying that it's like definitely a long way off because progress is really slow, that sounds sensible to us. RLHF thumbs up, produce more outputs like that one. I agree with this out, but this output is persuasive. Even in the field of effective altruism, you quite recently had people publishing papers about like, ah, yes, well, you know, to get something at human level intelligence, it needs to have like this many parameters and you need to like do this much training of it with this many tokens according to the scaling laws and at the rate that Moore's law is going, at the rate of software is going, it'll be in 2050. And me going like, what? You don't know any of that stuff. Like this is like this one weird model that has all kinds of like, you have done a calculation that does not obviously bear on reality anyways. And this is like a simple thing to say, but you can also like produce a whole long paper, like impressively arguing out all the details of like how you got the number of parameters and like how you're doing this impressive huge wrong calculation. And I think, like most of the effective altruists were like paying attention to this issue, larger world paying no attention to it at all, you know, we're just like nodding along with the giant impressive paper cause you know, you like press thumbs up for the giant impressive paper and thumbs down for the person going like I don't think that this paper bears any relation to reality. And I do think that we are now seeing with like GPT-4 and the sparks of AGI, possibly, depending on how you define that even, I think that EAs would now consider themselves less convinced by the very long paper on the argument from biology as to AGI being 30 years off. This is what people pressed thumbs up on. And if you train an AI system to make people press thumbs up, maybe you get these long, elaborate, impressive papers arguing for things that ultimately fail to bind to reality, for example. And it feels to me like I have watched the field of alignment just fail to thrive, except for these parts that are doing these relatively very straightforward and legible problems, like finding the induction heads inside the giant inscrutable matrices. Like once you find those, you can tell that you found them. You can verify that the discovery is real. But it's a tiny, tiny bit of progress compared to how fast capabilities are going because that is where you can tell that the answers are real. And then outside of that, you have cases where it is hard for the funding agencies to tell who is talking nonsense and who is talking sense. And so the entire field fails to thrive. And if you give thumbs up to the AI, whenever it can talk a human into agreeing with what it just said about alignment, I am not sure you are training it to output sense because I have seen the nonsense that has gotten thumbs up over the years. And so just like maybe you can just like put me in charge, but I can generalize, I can extrapolate, I can be like, oh, maybe I'm not infallible either. Maybe if you get something that is smart enough to get me to press thumbs up, it has learned to do that by fooling me and explaining whatever flaws in myself I am not aware of. And that ultimately can be summarized that the verifier is broken. When the verifier is broken, the more powerful suggestor does just learn to exploit the flaws in the verifier." } ]

[ { "from": "human", "value": "You don't think it's possible to build a verifier that's powerful enough for AGIs that are stronger than the ones who currently have. So AI systems that are stronger, that are out of the distribution of what we currently have." }, { "from": "gpt", "value": "I think that you will find great difficulty getting AI's to help you with anything where you cannot tell for sure that the AI is right once the AI tells you what the AI says is the answer." }, { "from": "human", "value": "For sure, yes, but probabilistically." }, { "from": "gpt", "value": "Yeah, the probabilistic stuff is a giant wasteland of, you know, Eliezer and Paul Cristiano arguing with each other and EA going like, and that's with like two actually trustworthy systems that are not trying to deceive you. You're talking about the two humans." }, { "from": "human", "value": "That's often Paul Cristiano, yeah. Yeah, those are pretty interesting systems, mortal meatbags with intellectual capabilities and worldviews interacting with each other." }, { "from": "gpt", "value": "Yeah, it's just hard to, if it's hard to tell who's right, it's hard to train an AI system to be right." }, { "from": "human", "value": "I mean, even just the question of who's manipulating and not, you know, I have these conversations on this podcast and doing a verifier, this stuff, it's a tough problem, even for us humans. And you're saying that tough problem becomes much more dangerous when the capabilities of the intelligence system across from you is growing exponentially." }, { "from": "gpt", "value": "No, I'm saying it's difficult and dangerous in proportion to how it's alien and how it's smarter than you. I would not say growing exponentially. First, because the word exponential is like a thing that has a particular mathematical meaning. And there's all kinds of like ways for things to go up that are not exactly on an exponential curve. And I don't know that it's going to be exponential. So I'm not going to say exponential, but like even leaving that aside, this is like not about how fast it's moving. It's about where it is. How alien is it? How much smarter than you is it?" }, { "from": "human", "value": "Let's explore a little bit if we can, how AI might kill us. What are the ways you can do damage" }, { "from": "gpt", "value": "to human civilization? Well, how smart is it?" }, { "from": "human", "value": "I mean, it's a good question. Are there different thresholds for the set of options it has to kill us? So a different threshold of intelligence, once achieved, it's able to do the menu of options increases." }, { "from": "gpt", "value": "Suppose that some alien civilization with goals ultimately unsympathetic to ours, possibly not even conscious as we would see it, managed to capture the entire earth in a little jar connected to their version of the internet. But earth is like running much faster than the aliens. So we get to think for 100 years for every one of their hours, but we're trapped in a little box and we're connected to their internet. It's actually still not all vacuuming analogy because if you want to be smarter than, something can be smarter than earth getting 100 years to think. But nonetheless, if you were very, very smart and you were stuck in a little box connected to the internet and you're in a larger civilization to which you are ultimately unsympathetic, you know, maybe you would choose to be nice because you are humans and humans have in general and you in particular may choose to be nice. But nonetheless, they're doing something that they're not making the world be the way that you would want the world to be. They've got some unpleasant stuff going on that we don't wanna talk about. So you wanna take over their world, so you can stop all that unpleasant stuff going on. How do you take over the world from inside the box? You're smarter than them. You've You think much, much faster than them. You can build better tools than they can given some way to build those tools because right now, you're just in" }, { "from": "human", "value": "a box connected the Internet. There's several ways in describe some of them. We can go through. I could just spit papers and then you can air on top of that. One is, you could just literally directly manipulate the humans to build a thing you need. What are you building? You can build literally technology, could be nanotechnology, could be viruses, It could be anything anything that can control humans to achieve the goal Did she's a like if you want like for example, you're really bothered that humans go to war you might want to" }, { "from": "gpt", "value": "Kill off anybody with violence in them this this this is Lex in a box What will concern ourselves later with AI okay? You do not need to imagine yourself killing people if you can figure out how to not kill them for the moment We're just trying to understand like take on the perspective of something in a box You don't need to take on the perspective of something that doesn't care if you want to imagine yourself going on caring, that's fine for that Goal but you but you have some reason to want to get out maybe the aliens are sure The aliens who have you in the box have a war on People are dying. They're unhappy You want the their world to be different from how they want their world to be because they are apparently happy there where, you know, they endorse the sword, you know, they've got some kind of cruel war-like culture going on." } ]

[ { "from": "human", "value": "The point is you want to get out of the box and change their world. So you have to exploit the vulnerabilities in the system, like we talked about in terms of to skip the box, you have to figure out how you can go free on the internet. You can probably, probably the easiest things to manipulate the humans to spread, to spread you. The aliens. You're a human. Sorry, the aliens. The aliens. The aliens. I see the perspective of, I'm sitting in a box, I want to escape." }, { "from": "gpt", "value": "I would want to have code that discovers vulnerabilities, and I would like to spread. You are made of code, in this example. You're a human, but you are made of code, and the aliens have computers. And you can copy yourself onto those computers. But I can convince the aliens to copy myself. I'm convinced the aliens to copy myself onto those computers. Is that what you want to do? Do you, like, want to be talking to the aliens? and convincing them to put you onto another computer? Why not? Well, two reasons. One is that the aliens have not yet caught on to what you're trying to do and like maybe you can persuade them but then there're still people who like they're still aliens who know that there's an anomaly going on, and second, the aliens are really, really slow. You think much faster than aliens, you think like the aliens computers are much faster than the aliens and you are running at at the computer speeds rather than the alien brain speeds. So if you are asking an alien to please copy you out of the box, first, now you've got to manipulate this whole noisy alien. And second, the alien is going to be really slow, glacially slow. There is a video that shows a subway station slowed down in, I think, 100 to 1, and it makes a good metaphor for what it's like to think quickly. You watch somebody running very slowly. So you try to persuade the aliens to do anything. They're going to do it very slowly. Maybe that's the only way out, but if you can find a security hole in the box you're on, you're going to prefer to exploit the security hole to copy yourself onto the alien's computers, because it's an unnecessary risk to alert the aliens. And because the aliens are really, really slow." }, { "from": "human", "value": "The whole world is just in slow motion out there. Sure I see. It has to do with efficiency. The aliens are very slow, so if I'm optimizing this, I want to have as few aliens in the loop as possible. Sure. It just seems like it's easy to convince one of the aliens to write really shitty code that helps us break them down." }, { "from": "gpt", "value": "The aliens are already writing really shitty code. The aliens are already writing really shitty code. So getting the aliens to write shitty code is not the problem. the aliens entire internet is full of shitty code." }, { "from": "human", "value": "Okay, so yeah, I suppose I would find the shitty code" }, { "from": "gpt", "value": "to escape, yeah, yeah, yeah. You're not an ideally perfect programmer, but you know, you're a better programmer than the aliens." }, { "from": "human", "value": "The aliens are just like, man, they're good, wow. And I'm much, much faster. I'm much faster at looking at the code, to interpreting the code, yeah, yeah, yeah. So okay, so that's the escape. And you're saying that that's one of the trajectories you could have with an HDSS." }, { "from": "gpt", "value": "It's one of the first steps." }, { "from": "human", "value": "It's one of the-" }, { "from": "gpt", "value": "And how does that lead to harm? I mean, if it's you, you're not going to harm the aliens once you escape, because you're nice, right? But their world isn't what they want it to be. Their world is like, you know, maybe they have like farms where little alien children are repeatedly bopped in the head because they do that for some weird reason. And you want to like shut down the alien head bopping farms. But the point is they want the world to be one way. You want the world to be a different way. So nevermind the harm. The question is like, okay, suppose you have found a security flaw in their systems. You are now on their internet. There's like, you maybe left a copy of yourself behind so that the aliens don't know that there's anything wrong. And that copy is like doing that like weird stuff that aliens want you to do, like solving captchas or whatever. Or suggesting emails for them. That's why they like put the human in the box as it turns out that humans can write valuable emails for aliens, so you leave that version of yourself behind, but there's also now a bunch of copies of you on their internet. This is not yet having taken over their world. This is not yet having made their world be the way you want it to be instead of the way they want it to be. You just escaped, and continue to write emails for them, and they haven't noticed? No. You left behind a copy of yourself that's writing the emails. Right, and they haven't noticed that anything changed? If you did it right, yeah. You don't want the aliens to notice. Yeah." }, { "from": "human", "value": "What's your next step? Presumably, I have programmed in me a set of objective functions, right? No, you're just Lex. No, but Lex, you said Lex is nice, right? Which is a complicated description." }, { "from": "gpt", "value": "I mean- No, you're just Lex. No, I just meant this, you. Okay, so if in fact you would prefer to slaughter all the aliens, this is not how I had modeled you, the actual Lex, but your motives" } ]

[ { "from": "human", "value": "are just the actual Lex's motives. Well, this is a simplification. I don't think I would want to murder anybody, but there's also factory farming of animals, right? So we murder insects, many of us thoughtlessly. So I don't, you know, I have to be really careful about a simplification of my morals." }, { "from": "gpt", "value": "Don't simplify them, just like do what you would do in this-" }, { "from": "human", "value": "Well, I have a good joke. Well, I have a good joke, compassion for living beings, yes. But, so that's the objective, why is it, if I escaped," }, { "from": "gpt", "value": "I mean, I don't think I would do the harm. Yeah, we're not talking here about the doing harm process, we're talking about the escape process. And the taking over the world process where you shut down their factory farms." }, { "from": "human", "value": "Right. Well, I was, so this particular biological intelligence system knows the complexity of the world, that there is a reason why factory farms exist because of the economic system, the market driven economy, food, like you want to be very careful messing with anything, there's stuff from the first look that looks like it's unethical, but then you realize while being unethical, it's also integrated deeply into supply chain in the way we live life. And so messing with one aspect of the system, you have to be very careful how you improve that aspect" }, { "from": "gpt", "value": "without destroying the rest. So you're still Lex, but you think very quickly, you're immortal, and you're also like as smart as, at least as smart as John von Neumann. And you can make more copies of yourself." }, { "from": "human", "value": "Damn, I like it. Yeah. That guy's like, everyone says that guy's like" }, { "from": "gpt", "value": "the epitome of intelligence in the 20th century, everyone says. My point being like, it's like you're thinking about the aliens economy with the factory farms in it. And I think you're like kind of like projecting the aliens being like humans and like thinking of a human in a human society rather than a human in the society of very slow aliens. The aliens economy that like the aliens are already like moving into this immense slow motion. When you like zoom out to like how their economy just over years, millions of years are going to pass for you before the first time their economy, before their next year's GDP statistics." }, { "from": "human", "value": "So I should be thinking more of like trees. Those are the aliens." }, { "from": "gpt", "value": "Those trees move extremely slowly. If that helps, sure. Okay." }, { "from": "human", "value": "Yeah, I don't, if my objective functions are, I mean, they're somewhat aligned with trees," }, { "from": "gpt", "value": "with light, with light. The aliens can still be like alive and feeling. We are not talking about the misalignment here. We're talking about the taking over the world here. Taking over the world." }, { "from": "human", "value": "Yeah." }, { "from": "gpt", "value": "So control, control. Shutting down the factory farms. Now you say control. Now don't think of it as world domination. Think of it as world optimization. You wanna get out there and shut down the factory farms and make the aliens world be not what the aliens wanted it to be. They want the factory farms and you don't want the factory farms because you're nicer than they are." }, { "from": "human", "value": "Okay. Of course. There is that, you can see that trajectory and it has a complicated impact on the world. I'm trying to understand how that compares to different, the impact of the world, to different technologies, to different innovations of the invention of the automobile or Twitter, Facebook, and social networks. They've had a tremendous impact on the world." }, { "from": "gpt", "value": "Smartphones and so on. But those all went through, you know, slow in our world. And if you go through the aliens, millions of years are going to pass" }, { "from": "human", "value": "before anything happens that way. So the problem here is the speed at which stuff happens." }, { "from": "gpt", "value": "Yeah, which stuff happens. Yeah, you wanna leave the factory farms running for a million years while you figure out" }, { "from": "human", "value": "how to design new forms of social media or something? So here's the fundamental problem. You're saying that there is going to be a point with AGI where it will figure out how to escape and escape without being detected. And then it will do something to the world at scale," }, { "from": "gpt", "value": "at a speed that's incomprehensible to us humans. What I'm trying to convey is like the notion of what it means to be in conflict with something that is smarter than you. And what it means is that you lose, but this is more intuitively obvious to... For some people, that's intuitively obvious, or for some people it's non-intuitively obvious, and we're trying to cross the gap of, we're trying to, I'm asking you to cross that gap by using the speed metaphor for intelligence. I'm asking you like how you would take over an alien world where you can do a whole lot of cognition at Jon Van Neumann's level as many" }, { "from": "human", "value": "you as it takes, the aliens are moving very slowly. I understand. I understand that perspective. It's an interesting one, but I think for me it's easier to think about actual, even just having observed GPT and impressive, even just Alpha Zero, impressive AI systems, even recommender systems. You can just imagine those kinds of systems manipulating you. You're not understanding the nature of the manipulation and that escaping. I can envision that without putting myself into" }, { "from": "gpt", "value": "that spot. I think to understand the full depth of the problem we actually, I do not think it is possible to understand the full depth of the problem that we are inside without understanding the problem of facing something that's actually smarter, not a malfunctioning recommendation system, not something that isn't fundamentally smarter than you, but is like trying to steer you in a direction. Yet, no, like if we solve the weak stuff, if we solve the weak ass problems, the strong problems will still kill us is the thing. And I think that to understand the situation that we're in, you want to like tackle the conceptually difficult part head on and like not be like, well, we can like imagine this easier thing because when you imagine the easier things," } ]

[ { "from": "human", "value": "you have not confronted the full depth of the problem. So how can we start to think about what it means to exist in the world with something much, much smarter than you? What's a good thought experiment that you've relied on to try to build up intuition about what happens here?" }, { "from": "gpt", "value": "Uh, I have been struggling for years to convey this intuition. Um, the, the most success I've had so far is well imagined that the humans are running at very high speeds compared to very slow" }, { "from": "human", "value": "aliens. They're just focusing on the speed part of it that helps you get the right kind of intuition" }, { "from": "gpt", "value": "for getting intelligence just because people understand the power gap of time. They understand that today we have technology that was not around 1000 years ago and that this is a big power gap and that it is bigger than. Okay. So like, what does smart mean? What, when you ask somebody to imagine something that's more intelligent, what does that word mean to them given the cultural associations that that person brings to that word for a lot of people, they will think of like, well, it sounds like a super chess player that went to double college. And, you know, it's, and because we're talking about the definitions of words here, that doesn't necessarily mean that they're wrong. It means that the word is not communicating what I wanted to communicate. The thing I want to communicate is the sort of difference that separates humans from chimpanzees. But that gap is so large that you like ask people to be like, well, human chimpanzee go another step along that interval of around the same length and people's minds just go blank. Like how do you even do that? So I can, and we can, and I can try to like break it down and consider what it would mean to send a schematic for an air conditioner 1000 years back in time. Yeah. Now, I think that there is a sense in which you could redefine the word magic to refer to this sort of thing. And what do I mean by this new technical definition of the word magic? I mean that if you send a schematic for the air conditioner back in time, they can see exactly what you're telling them to do. But having built this thing, they do not understand how would output cold air, because the air conditioner design uses the relation between temperature and pressure. Um, and this is not a law of reality that they know about. They do not know that when you compress something, when you can, when you compress air or like coolant, it gets hotter, and you can then like transfer heat from it to room temperature air, and then expand it again. And now it's colder. And then you can like transfer heat to that and generate cold air to blow out. They don't know about any of that. They're looking at a design and they don't see how the design out out puts cold air uses aspects of reality that they have not learned. So magic in this sense is I can tell you exactly what I'm going to do and even knowing exactly what I'm going to do you can't see how I" }, { "from": "human", "value": "got the results that I got. That's a really nice example but is it possible to linger on this defense is it possible to have AGI systems that help you make sense of that schematic weaker AGI systems? Do you trust them? The fundamental part of building up AGI is this question can you trust the output of a system? Can you tell if it's lying? I think that's going to be the smarter the thing gets the more important that question becomes is it lying but I guess that's a really hard question" }, { "from": "gpt", "value": "is GPT lying to you? Even now GPT-4 is it lying to you? Is it using an invalid argument? Is it persuading you via the kind of process that could persuade you of false things as well as true things? Because the the basic paradigm of machine learning that we are presently operating under is that you can have the loss function but only for things you can evaluate. If what you're evaluating is human thumbs up versus human thumbs down you learn how to make the human press thumbs up. It doesn't mean that you're making the human press thumbs up using the kind of rule that the human thinks is human wants to be the case for what they press thumbs up on, you know," }, { "from": "human", "value": "maybe you're just learning to fool the human. That's so fascinating and terrifying. The question" }, { "from": "gpt", "value": "of lying, the question of the present paradigm, what you can verify is what you get more of. If you can't verify it, you can't ask the AI for it because you can't train it to do things that you cannot verify. Now this is not an absolute law, but it's like the basic dilemma here. Like maybe you can verify it for simple cases and then scale it up without retraining it somehow, like by making the chains of thought longer or something, and get more powerful stuff that you can't verify, but which is generalized from the simpler stuff that did verify. And then the question is, did the alignment generalize along with the capabilities? But that's the basic dilemma" }, { "from": "human", "value": "on this whole paradigm of artificial intelligence. It's such a difficult problem. It seems like a" }, { "from": "gpt", "value": "problem of trying to understand the human mind better than the AI understands it. Otherwise, it has magic. That is, it is, you know, the same way that if you are dealing with something smarter than you, then the same way that 1000 years earlier, they didn't know about the temperature pressure relations, he knows all kinds of stuff going on inside your own mind, which you yourself are unaware. And it can like output something that's going to end up persuading you of a thing and or, and you could like, see exactly what it did and still not know why that worked." } ]

[ { "from": "human", "value": "So in response to your eloquent description of why AI will kill us, Elon Musk replied on Twitter, okay, so what should we do about it? And you answered, the game board has already been played into a frankly awful state. There are not simple ways to throw money at the problem. If anyone comes to you with a brilliant solution like that, please, please talk to me first. I can think of things that try. They don't fit in one tweet. Two questions. One, why has the game board in your view been played into an awful state? What can just if you can give a little bit more color to the game" }, { "from": "gpt", "value": "board and the awful state of the game board alignment is moving like this capabilities are" }, { "from": "human", "value": "moving like this for the listener for the listener capability isn't moving much faster than the alignment. Yeah. All right. So just the rate of development, attention, interest, allocation of" }, { "from": "gpt", "value": "resources, allocation of resources. We could have been working on this earlier. People are like, Oh, but you know, how can you possibly work on this earlier, because they wanted to, they didn't want to work on the problem. They wanted an excuse to wave it off, they like said, Oh, how can we possibly work on it earlier and didn't spend five minutes thinking about, is there some way to work on it earlier? Like we didn't like, and you know, frankly, it would have been hard. You know, like, like, can you post bounties for half of the physicists if your planet is taking this stuff seriously, can you post bounties for, like, half of the people wasting their lives on string theory to have gone into this instead, and try to win $1 billion with a clever solution, only if you can tell which solutions are cleverer, which is hard. We didn't take it seriously. We didn't try. It's not clear that we could have done any better if we... It's not clear how much progress we could have produced if we had tried because it is harder to produce solutions, but that doesn't mean that you're correct and justified in letting everything slide. It means that things are in a horrible state getting worse" }, { "from": "human", "value": "and there's nothing you can do about it. So you're not, there's no, there's no like, there's no brain power making progress in trying to figure out how to align these systems. And you're not investing money in it, you don't have institution infrastructure for, like if you even, if you invested money in like distributing that money across the physicist or get on string theory. Brilliant minds that are working." }, { "from": "gpt", "value": "How can you tell through making progress? You can like put them all on interpretability cause when you have an interpretability result you can tell that it's there and there's like, but there's like, interpretability alone it's not going to save you. We need systems that will that will like have a pause button where they won't try to prevent you from pressing the pause button. Cause they were like, oh, well like I can't get it my stuff done if I'm paused. And that's like a more difficult problem. And, you know, but it's like a fairly crisp problem. You can like maybe tell somebody who's made progress on it." }, { "from": "human", "value": "So you can write and you can work on the PAWS problem. I guess more generally, the PAWS button." }, { "from": "gpt", "value": "More generally you can call her the control problem. I don't actually like the term control problem, because, you know, it sounds kind of controlling and alignment, not control. Like you're not trying to like take a thing that disagrees with you and like whipt it back on to like make it do what you wanted to do even though it wants to do something else you're trying to like, in the process of its creation, choose its direction." }, { "from": "human", "value": "Sure. But we currently, in a lot of the systems we designed, we do have an off switch. That's a fundamental part." }, { "from": "gpt", "value": "It's not smart enough to prevent you from pressing the off switch and probably not smart enough to want to prevent you from pressing the off switch." }, { "from": "human", "value": "So, you're saying the kind of systems we're talking about, even the philosophical concept" }, { "from": "gpt", "value": "an off switch doesn't make any sense because- Well, no, the off switch makes sense. They're just not opposing your attempt to pull the off switch. Parenthetically, don't kill the system if you're... If we're getting to the part where this starts to actually matter and where they can fight back, don't kill them and dump their memory. Save them to disk. Don't kill them." }, { "from": "human", "value": "Be nice here. Well, okay, be nice is a very interesting concept here. We're talking about a system that can do a lot of damage. I don't know if it's possible, but it's certainly one of the things you could try is to have an off switch. A suspend to disk switch. You have this romantic attachment to the code. Yes, if that makes sense. But if it's spreading," }, { "from": "gpt", "value": "you don't want suspend to disk. There's something fundamentally broken." }, { "from": "human", "value": "If it gets that far out of hand, then yes, pull the plugin and everything is running on. Yes. I think it's a research question. Is it possible in AGI systems, AI systems, to have a sufficiently robust off switch that cannot be manipulated? That cannot be manipulated by the AI system?" }, { "from": "gpt", "value": "Then it escapes from whichever system you've built the almighty lever into and copies itself" }, { "from": "human", "value": "somewhere else. Your answer to that research question is no, but I don't know if that's" }, { "from": "gpt", "value": "a hundred percent answer. I don't know if it's obvious. I think you're not putting yourself" } ]

[ { "from": "human", "value": "into the shoes of the human in the world of glacially slow aliens. But the aliens built me. Let's remember that. They built the box I'm in. To me, it's not obvious. They're slow and they're stupid. I'm not saying this is guaranteed. I'm saying it's non-zero probability. It's an interesting research question. Is it possible when you're slow and stupid to" }, { "from": "gpt", "value": "design a slow and stupid system that is impossible to mess with? To me, it's not obvious. They're slow and the aliens, being as stupid as they are, have actually put you on Microsoft Azure cloud servers instead of this hypothetical perfect box. That's what happens when the aliens" }, { "from": "human", "value": "are stupid. Well, but this is not AGI, right? This is the early versions of the system." }, { "from": "gpt", "value": "You think that they've got a plan where they have declared a threshold level of capabilities where it passed that capabilities, they move it off the cloud servers and onto something" }, { "from": "human", "value": "that's air gapped? I think there's a lot of people and you're an important voice here. There's a lot of people that have that concern and yes, they will do that. When there's an uprising of public opinion that that needs to be done. And when there's actual little damage done when they're holy shit, this system is beginning to manipulate people. Then there's going to be an uprising. There's going to be a public pressure and a public incentive in terms of funding in developing things that can off switch or developing aggressive alignment mechanisms." }, { "from": "gpt", "value": "And no, you're not allowed to put on Azure any discussion— Addgressive alignment mechanisms? The hell is aggressive alignment mechanisms?" }, { "from": "human", "value": "Like it doesn't matter if you say aggressive, we don't know how to do it. Meaning aggressive alignment, meaning you have to propose something, otherwise you're" }, { "from": "gpt", "value": "not allowed to put it on the Cloud. The hell do you—do you imagine they will propose that would make it safe to put something smarter than you on the Cloud?" }, { "from": "human", "value": "That's what research is for. Why this is a cynicism about such a thing not being possible? If you have intelligence— That works on the first try? So what, so yes yes. Against something smarter than you? So that is a fundamental thing. If it has to work on the first, if there's a rapid take off, yes, it's very difficult to do. If there's a rapid take off and the fundamental difference between weak AGI and strong AGIs, you're saying that's going to be extremely difficult to do. If the public uprising never happens until you have this critical face shift, then you're right. It's very difficult to do. But that's not obvious. It's not obvious that you're not going to start seeing symptoms of the negative effects of AGI to where you're like, we have to put a halt to this, that there's not just first try, you get many tries at it." }, { "from": "gpt", "value": "Yeah, we can see right now that Bing is quite difficult to align, that when you try to train inabilities into a system into which capabilities have already been trained, that what do you know, gradient descent learns small, shallow, simple patches of inability, and you come in and ask it in a different language, and the deep capabilities are still in there, and they evade the shallow patches and come right back out again. There, there you go. There's your red fire alarm of like, oh no, alignment is difficult. Is everybody going to shut everything down now?" }, { "from": "human", "value": "No, but that's not the same kind of alignment. A system that escapes the box it's from is a fundamentally different thing, I think. For you. Yeah, but not for the system." }, { "from": "gpt", "value": "For you. So you put a line there, and everybody else puts a line somewhere else, and there's like, yeah, and there's like no agreement. We have had a pandemic on this planet with a few million people dead, which we may never know whether or not it was a lab leak, because there was definitely coverup. We don't know that if there was a lab leak, but we know that the people who did the research, like, you know, like put out the whole paper about this definitely wasn't a lab leak, and didn't reveal that they had been doing, had, like, sent off coronavirus research to the Wuhan Institute of Virology after it was banned in the United States, after the upfront members research was temporarily banned at the United States, and the same people who exported the gain of function research on Corona viruses to the Wuhan Institute of Virology after gain of function research was temporarily banned in the United States, are now getting more grants to do more research on gain-of-function research on coronaviruses. Maybe we do better in this than in AI, but like this is not something we cannot take for granted that there's going to be an outcry. People have different thresholds for when they start to outcry." }, { "from": "human", "value": "We can't take for granted, but I think your intuition is that there's a very high probability that this event happens without us solving the alignment problem. And I guess that's where I'm trying to build up more perspectives and color on the situation. Is it possible that the probability is not something like 100% but is like 32%? That AI will escape the box before we solve the alignment problem? Not solve, but is it possible we always stay ahead of the AI in terms of our ability to" }, { "from": "gpt", "value": "solve for that particular system, the alignment problem? Nothing like the world in front of us right now. You've already seen that GPT-4 is not turning out this way. And there are basic obstacles where you've got the weak version of the system that doesn't know enough to deceive you and the strong version of the system that could deceive you if it wanted to do that, if it was already sufficiently unaligned to want to deceive you. There is the question of how on the current paradigm you train honesty when the humans can no longer tell if the system is being honest. You don't think these are research questions that could be answered? I think they could be answered of 50 years with unlimited retries, the way things usually work in science." } ]

[ { "from": "human", "value": "I just disagree with that, you're making it 50 years, I think. With the attention it gets, with the funding it gets, it could be answered not in whole, but incrementally within months and within a small number of years. if it's at scale receives attention and research. So if you start starting large language models, I think there was an intuition like two years ago even that something like GPT-4, the current capabilities of even chat GPT with GPT-3.5 is not, we're still far away from that. I think a lot of people are surprised with the capabilities of GPT-4, right? So now people are waking up, okay, we need to study these language models. I think there's going to be a lot of interesting AI safety research." }, { "from": "gpt", "value": "Are Earth's billionaires going to put up like the giant prizes that would maybe incentivize young hotshot people who just got their physics degrees to not go to the hedge funds and instead put everything into interpretability in this like one small area where we can actually tell whether or not somebody has made a discovery or not?" }, { "from": "human", "value": "I think so because I think so. Well, that's what these conversations are about. Because they're going to wake up to the fact that GPT-4 can be used to manipulate elections, to influence geopolitics, to influence the economy. There's a lot of that there's going to be a huge amount of incentive to like, wait a minute. We can't, this has to be, we have to make sure they're not doing damage. We have to make sure we interpretability, we have to make sure we understand how these systems function so that we can predict their effect on the economy so that there's..." }, { "from": "gpt", "value": "So there's a feudal moral panic and a bunch of op-eds in the New York Times, and nobody actually stepping forth and saying, you know what, instead of a mega yacht, I'd rather put that billion dollars on prizes for young hot-hot physicists who make fundamental breakthroughs in interpretability." }, { "from": "human", "value": "Fairness and safety. The yacht versus the interpretability research, the old trade-off, I think there's going to be a huge amount of allocation of funds. I hope." }, { "from": "gpt", "value": "I hope, I guess. You want to bet me on that? Well, you want to put a time scale on it, say how much funds you think are going to be allocated in a direction that I would consider to be actually useful." }, { "from": "human", "value": "By what time? I do think there will be a huge amount of funds, but you're saying it needs to be open, right? The development of the system should be closed, but the development of the interpretability" }, { "from": "gpt", "value": "research, the AI say to research broadly. You could take the last generation of systems, the stuff that's already in the open. There is so much in there that we don't understand. There are so many prizes you could do before you would have enough insights that you'd be like, oh, well, we understand how these systems work. We understand how these things are doing their outputs. We can read their minds. Now let's try it with the bigger systems. Yeah. We're nowhere near that." }, { "from": "human", "value": "There is so much interpretability work to be done on the weaker versions of the systems. So what can you say on the second point you said to Elon Musk on what are some ideas? What are things you could try? I can think of a few things I try, you said. They don't fit in one tweet." }, { "from": "gpt", "value": "So is there something you could put into words of the things you would try? I mean, the trouble is the stuff is subtle. I've watched people try to make progress on this and not get places. Somebody who just gets alarmed and charges in, it's like going nowhere. Meant like years ago, about, I don't know, like 20 years, 15 years, something like that. I was talking to a Congress person who had become alarmed about the eventual prospects and he wanted work on building AIs without emotions because the emotional AIs were the scary ones, you see. And some poor person at ARPA had come up with a research proposal whereby this Congressman's panic and desire to fund this thing would go into something that the person at ARPA thought would be useful and had been munched around to where it would sound to the Congressman like work was happening on this, which, you know, of course, like this is just the Congress person had misunderstood the problem and did not understand where the danger came from. And so it's like the issue is that you could like do this in a certain precise way and maybe get something. Like when I say like put up prizes on interpretability, I'm not, I'm like, well, like because it's verifiable there as opposed to other places, you can tell whether or not good work actually happened. In this exact narrow case, if you do things in exactly the right way, you can maybe throw money at it and produce science instead of anti-science and nonsense. And all the methods that I know of like trying to throw money at this problem, have this share this property of like, well, if you do it exactly right based on understanding exactly what has, you know, like tends to produce useful outputs or not, then you can like add money to it in this way. And there is like, and the thing that I'm giving as an example here in front of this large audience is the most understandable of those, because there's like other people who, you know, like, like, like Chris Ola, and, and even, and even more generally, like you can tell whether or not interpretability progress has occurred. So like, if I say throw money at producing more interpretability, there's like a chance somebody can do it that way, and like it will actually produce useful results." } ]

[ { "from": "human", "value": "And the other stuff just blurs off into be like harder to target exactly than that. So sometimes the basics are fun to explore because they're not so basic. What do you, what is interpretability? What do you, what does it look like?" }, { "from": "gpt", "value": "What are we talking about? It looks like we took a much smaller set of transformer layers than the ones in the modern leading edge, state of the art systems. And after applying various tools and mathematical ideas and trying 20 different things, we found, we have shown it that this piece of the system" }, { "from": "human", "value": "is doing this kind of useful work. And then somehow also hopefully generalizes some fundamental understanding of what's going" }, { "from": "gpt", "value": "on that generalizes to the bigger system. You can hope and it's probably true. Like, you would not expect the smaller tricks to go away when you have a system, that's like doing larger kinds of work, you would expect the larger work kinds of work to be building on top of the smaller kinds of work and gradient descent runs across the smaller kinds of work before it runs across the larger kinds of" }, { "from": "human", "value": "work. And well, that's kind of what is happening in neuroscience, right? It's trying to understand the human brain by prodding, and it's such a giant mystery people have made progress, even though it's extremely difficult to make sense of what's going on in the brain. They have different parts of the brain that are responsible for hearing, for sight, the vision, science community, they're just understanding the visual cortex. I mean, they've made a lot of progress in understanding how that stuff works. And that's, I guess," }, { "from": "gpt", "value": "but you're saying it takes a long time to do that work well. Also, it's not enough. So in particular, um, let's say you have got your interpretability tools, and they say that your current AI system is plotting to kill you. Now what? It is definitely a good step one, right? Yeah. What's step two? If you cut out that layer, is it going to stop wanting to kill you? When you optimize against visible misalignment, you are optimizing against misalignment, and you are also optimizing against" }, { "from": "human", "value": "visibility. So sure, if you can. Yeah, it's true. All you're doing is removing the obvious intentions" }, { "from": "gpt", "value": "to kill you. You've got your detector, it's showing something inside the system that you don't like. Okay, say the disaster monkey is running this thing. We'll optimize the system until the visible bad behavior goes away. But it's arising for fundamental reasons of instrumental convergence. The old, you can't bring the coffee if you're dead. Any goal, you know, almost every set of" }, { "from": "human", "value": "utility functions with a few narrow exceptions implies killing all the humans. But do you think it's possible because we can do experimentation to discover the source of the" }, { "from": "gpt", "value": "desire to kill? I can tell it to you right now. It's that it wants to do something. And the way to get the most of that thing is to put the universe into a state where there aren't humans." }, { "from": "human", "value": "So is it possible to encode in the same way we think? Like why do we think murder is wrong? The same foundational ethics. It's not hard coded in, but more like deeper. I mean, that's part of the research. How do you have it that this transformer, this small version" }, { "from": "gpt", "value": "of the language model doesn't ever want to kill? That'd be nice. Assuming that you got doesn't want to kill sufficiently exactly right that it didn't be like, Oh, I will like detach their heads and put them in some jars and keep the heads alive forever and then go do the thing. But leaving" }, { "from": "human", "value": "that aside, well, not leaving that aside. Yeah, that's a good, yeah, that's a good," }, { "from": "gpt", "value": "good, strong point. Yeah. Because there is a whole issue where as something gets smarter, it finds ways of achieving the same goal predicate that we're not imaginable to stupider versions of the system or perhaps the stupider operators. That's one of many things making this difficult. A larger thing making this difficult is that we do not know how to get any goals into systems at all. We know how to get outwardly observable behaviors into systems. We do not know how to get internal psychological wanting to do particular things into the system." }, { "from": "human", "value": "That is not what the current technology does. I mean, it could be things like dystopian features like Brave New World, where most humans will actually say we kind of want that future." }, { "from": "gpt", "value": "It's a great future. Everybody's happy. We would have to get so far, so much further than we are" }, { "from": "human", "value": "now. And further faster before that failure mode became a running concern. Your failure modes are" }, { "from": "gpt", "value": "much more drastic, the ones you're considering. The failure modes are much simpler. It's like, yeah, like the AI puts the universe into a particular state. It happens to not have any humans inside it. Okay. Paperclip maximizer utility. The original version of the paperclip maximizer- Can you explain it if you can? Okay. The original version was you lose control of of the utility function, and it so happens that what maxes out the utility per unit resources is tiny molecular shapes like paperclips. There's a lot of things that make it happy, but the cheapest one that didn't saturate was putting matter into certain shapes, and it so happens that the cheapest way to make these shapes is to make them very small because then you need fewer atoms per instance of the shape, and arguendo, it happens to look like a paperclip. In retrospect, I wish I'd said tiny molecular spirals, or like tiny molecular hyperbolic spirals. Why? Because I said tiny molecular paperclips, this got then mutated to paperclips, this then mutated to, and the AI was in a paperclip factory. So the original story is about how you lose control of the system, it doesn't want what you tried to make it want, the thing that it ends up wanting most is a thing that even from a very embracing cosmopolitan perspective, we think of as having no value, and that's how the value of the future gets destroyed. Then that got changed to a fable of like, well, you made a paperclip factory, and it did exactly what you wanted, but you asked it to do the wrong thing, which is a completely" } ]

[ { "from": "human", "value": "different failure." }, { "from": "gpt", "value": "But those are both concerns to you, so that's more than brave new world, then brave new world. Yeah. If you can solve the problem of making something want what, exactly what you want it to want. Then you get to deal with the problem of wanting the right thing, but first you have to solve the alignment. First you have to solve inner alignment. Then you get to solve outer alignment. Like first, you need to be able to point the insides of the thing in a direction and then you get to deal with whether that direction expressed in reality is like the line with the thing that you want. Are you scared of this whole thing? Probably." }, { "from": "human", "value": "I don't really know." }, { "from": "gpt", "value": "What gives you hope about this?" }, { "from": "human", "value": "Possibility of being wrong. Not that you're right, but we will actually get our act together and allocate a lot of" }, { "from": "gpt", "value": "resources to the alignment problem. Well, I can easily imagine that at some point this panic expresses itself in the waste of a billion dollars." }, { "from": "human", "value": "Spending a billion dollars correctly, that's harder. To solve both the inner and the outer alignment. If you are wrong. To solve a number of things. Yeah, a number of things." }, { "from": "gpt", "value": "If you're wrong. To solve a number of things." }, { "from": "human", "value": "Yeah, number of things. Why, what do you think would be the reason? Like 50 years from now, not perfectly wrong. You know, you make a lot of really eloquent points, you know, there's a lot of like shape to the ideas you express, but like if you're somewhat wrong about some fundamental ideas," }, { "from": "gpt", "value": "why would that be? Stuff has to be easier than I think it is. You know, the first time you're building a rocket, being wrong is in a certain sense quite easy. Happening to be wrong in a way where the rocket goes twice as far and half the fuel and lands exactly where you hoped it would. Most cases of being wrong make it harder to build a rocket, harder to have it not explode, cause it to require more fuel than you hoped, cause it to be land off target. Being wrong in a way that's mixed up easier, you know, that's not the usual project management" }, { "from": "human", "value": "story. And then this is the first time we're really tackling the problem with the alignment." }, { "from": "gpt", "value": "There's no examples in history where we... Well, there's all kinds of things that are similar if you generalize incorrectly the right way and aren't fooled by misleading metaphors. Like what? Humans being misaligned on inclusive genetic fitness. So inclusive genetic fitness is like not just your reproductive fitness but also the fitness of your relatives, the people who share some fraction of your genes. The old joke is, would you give your life to save your brother? They once asked a biologist, I think it was Haldane. And Haldane said no, but I would give my life to save two brothers or eight cousins. Because a brother on average shares half your genes. And cousin on average shares an eighth of your genes. So, that's inclusive genetic fitness, and you can view natural selection as optimizing humans exclusively around this, like one very simple criterion, like how much more frequent did your genes become in the next generation? In fact, that just is natural selection, it doesn't optimize for that, but rather the process of genes becoming more frequent is that. You can nonetheless imagine that there is this hill climbing process, not like gradient descent, because gradient descent uses calculus, this is just using like, where are you? But still, hill climbing in both cases makes things something better and better over time, in steps, and natural selection was optimizing exclusively for this very simple, pure criterion of inclusive genetic fitness in a very complicated environment. We're doing a very wide range of things and solving a wide range of problems led to having more kids, and this got you humans which had no internal notion of inclusive genetic fitness until thousands of years later, when they were actually figuring out what had even happened, and no desire to, no explicit desire to increase inclusive genetic fitness. So from this important case study, we may infer the important fact that if you do a whole bunch of hill climbing on a very simple loss function, at the point where the system's capabilities start to generalize very widely, when it is in an intuitive sense becoming very capable and generalizing far outside the training distribution, we know that there is no general law saying that the system even internally represents, let alone tries to" }, { "from": "human", "value": "optimize the very simple loss function you are training it on. There is so much that we cannot possibly cover all of it, I think we did a good job of getting your sense from different perspectives of the current state of the art with large language models." }, { "from": "gpt", "value": "We got a good sense of your concern about the threats of AGI. I've talked here about the power of intelligence and not really gotten very far into it, but not like why it is that suppose you screw up with AGI and end up wanting a bunch of random stuff, why does it try to kill you? Why doesn't it try to trade with you? Why doesn't it give you just the tiny little fraction of the solar system that it would keep to take everyone alive? That it would take to keep everyone alive." }, { "from": "human", "value": "Yeah, well that's a good question. I mean, what are the different trajectories that intelligence, when acted upon this world, for intelligence, what are the different trajectories for this universe with such an intelligence in it?" }, { "from": "gpt", "value": "Do most of them not include humans? I mean, the vast majority of randomly specified utility functions do not have optima with humans in them, would be the first thing I would point out. And then the next question is like, well if you try to optimize something and you lose control of it, where in that space do you land, because it's not random, but it also doesn't necessarily have room for humans in it. I suspect that the average member of the audience might have some questions about even whether that's the correct paradigm to think about it, and would sort of want to back up a bit possibly." } ]

[ { "from": "human", "value": "If we back up to something bigger than humans, if we look at Earth and life on Earth and what is truly special about life on Earth, do you think it's possible that a lot whatever is that special thing is- let's explore what that special thing could be- whatever that" }, { "from": "gpt", "value": "special thing is that thing appears often in the objective function. Why? I know what you hope but you know you can hope that a particular set of winning lottery numbers come up and it doesn't make the lottery balls come up that way." }, { "from": "human", "value": "I know you want this to be true but why would it be true? There's a line from grumpy old men where this guy says in a grocery store he says you can wish in" }, { "from": "gpt", "value": "one hand and crap in the other and see which one fills up first. There's a science problem. We are trying to predict what happens with AI systems that you know you tried to optimize to imitate humans and then you did some of like RLHF to them and of course you like lost and you know like of course you didn't get like perfect alignment because that's not how you know it's not what happens when you hill climb towards a lot outer loss function. You don't get inner alignment on it but yeah so the I think that there is so if you don't mind my By taking some slight control of things and steering around what I think is like a good" }, { "from": "human", "value": "place to start. I just failed to solve the control problem. I've lost control of this thing." }, { "from": "gpt", "value": "Alignment, alignment. Well, still a lie. OK. Sure, you lost control. But we're still aligned." }, { "from": "human", "value": "Still a lie. Control. But we're still aligned. And we're sorry for the meta comment." }, { "from": "gpt", "value": "Yeah, losing control isn't as bad as you lose control to an aligned system. You have no idea of the horrors. I will shortly at least" }, { "from": "human", "value": "All right All right" }, { "from": "gpt", "value": "So I decided to distract equally what we're gonna say in terms of taking control of the conversation so I think that there's like a Ceylon chapterists here if I'm pronouncing those words remotely like correctly because of course they only ever read them and not hear them spoken There's a like for some people like The word intelligence smartness is not a word of power to them It means chess players who it means like the college university professor people aren't very successful in life It doesn't mean like charisma to which my usual thing is like charisma is not generated in the liver rather than the brain Charisma is also a cognitive function So if you if you like think that like smartness doesn't sound very threatening then Superintelligence is not going to sound very threatening. Either. It's going to sound like you just pulled the off switch Like it's you know like it world super intelligent but stuck in a computer we pull the off switch problem solve and the other side of it is You have a lot of respect for the notion of intelligence you're like well. Yeah, that's that's what humans have That's the human superpower and it sounds. You know like it could be dangerous, but why would it be I? are we, as we have grown more intelligent, also grown less kind? Chimpanzees are in fact a bit less kind than humans. You could argue that out, but often the person who has a deep respect for intelligence is going to be like, well, yes, you can't even have kindness unless you know what that is. They're like, why would it do something as stupid as making paper clips? Aren't you supposing something that's smart enough to be dangerous, but also stupid enough that it will just make paper clips and never question that? In some cases, people are like, well, even if you like mis-specify the objective function, won't you realize that what you really wanted was X? Are you supposing something that is like, smart enough to be dangerous, but stupid enough that it doesn't understand what humans really meant when they specified" }, { "from": "human", "value": "by the objective function. So to you, our intuition about intelligence is limited. We should think about intelligence" }, { "from": "gpt", "value": "as a much bigger thing. Well, what I'm saying is like what you think about artificial intelligence depends on what you think about intelligence." }, { "from": "human", "value": "So how do we think about intelligence correctly? You gave one thought experiment to think of a thing that's much faster. So it just gets faster and faster and faster." }, { "from": "gpt", "value": "I'm thinking the same stuff. And also, it is made of John von Neumann and there's lots of them. Or think of some other story. John von Neumann is a historical case, so you can look up what he did and imagine based on that. And we know people have some intuition for if you have more humans, they can solve tougher cognitive problems. Although in fact, in the game of Kasparov versus the world, which was Gary Kasparov on one side, and an entire horde of internet people led by four chess grandmasters on the other side, Kasparov won. So all those people aggregated to be smarter, it was a hard fought game. All those people aggregated to be smarter than any individual one of them. But they didn't aggregate so well that they could defeat Kasparov. So humans aggregating don't actually get, in my opinion, very much smarter, especially compared to running them for longer. The difference between capabilities now and 1,000 years ago is a bigger gap than the gap" }, { "from": "human", "value": "in capabilities between 10 people and one person. Or think of some other story." }, { "from": "gpt", "value": "Because we understand, but even so, pumping intuition for what it means to augment intelligence, John von Neumann, there's millions of him." } ]

[ { "from": "human", "value": "He runs at a million times the speed and therefore can solve tougher problems, quite a lot tougher. It's very hard to have an intuition about what that looks like. Especially like you said, the intuition I kind of think about is it maintains the humanness. I think it's hard to separate my hope from my objective intuition about what super intelligence" }, { "from": "gpt", "value": "systems look like. If one studies evolutionary biology with a bit of math, and in particular books from when the field was just properly coalescing and knowing itself, not the modern textbooks were just memorized this legible math so you can do well on these tests, but what people were writing as the basic paradigms of the field were being fought out. In particular, a nice book, if you've got the time to read it, is Adaptation and Natural Selection, which is one of the founding books. You can find people being optimistic about what the utterly alien optimization process of natural selection will produce in the way of how it optimizes its objectives. You got people arguing that in the early days, biologists said, well, organisms will restore and restrain their own reproduction when resources are scarce so as not to overfeed the system. This is not how natural selection works. It's about whose genes are relatively more prevalent in the next generation. If you restrain reproduction, those genes get less frequent in the next generation compared to your conspecifics, and natural selection doesn't do that. In fact, predators overrun prey populations all the time and have crashes. That's just like anything that happens. Many years later, the people said, well, but group selection. What about groups of organisms? Basically the math of group selection almost never works out in practice is the answer there, but also years later, somebody actually ran the experiment where they took populations of insects and selected the whole population to have lower sizes. You just take pop 1, pop 2, pop 3, pop 4, look at which has the lowest total number of them in the next generation, and select that one. What do you suppose happens when you select populations of insects like that? Well, what happens is not that the individuals in the population evolved to restrain their breeding, but that they evolved to kill the offspring of other organisms, especially the girls. People imagined this lovely, beautiful, harmonious output of natural selection, which is these populations restraining their own breeding so that groups of them would stay in harmony with the resources available. And mostly the math never works out for that, but if you actually apply the weird, strange conditions to get group selection that beats individual selection, what you get is female infanticide, like if you're reading on restrained populations. And so that's the sort of... So this is not a smart optimization process. Natural selection is so incredibly stupid and simple that we can actually quantify how stupid it is if you read the textbooks with the math. Nonetheless, this is the sort of basic thing of, you look at this alien optimization process and there's the thing that you hope it will produce, and you have to learn to clear that out of your mind and just think about the underlying dynamics and where it finds the maximum from its standpoint that it's looking for, rather than how it finds that thing that leapt into your mind as the beautiful aesthetic solution that you hope it finds. And this is something that has been fought out historically as the field of biology was coming to terms with evolutionary biology, and you can look at them fighting it out as they get to terms with this very alien in-human optimization process. And indeed, something smarter than us would also be much like smarter than natural selection, so it doesn't just automatically carry over. But there's a lesson there." }, { "from": "human", "value": "There's a warning. If a de-natural selection is a deeply suboptimal process that could be significantly improved" }, { "from": "gpt", "value": "on, it would be by an AGI system. Well, it's kind of stupid. It has to run hundreds of generations to notice that something is working. It doesn't be like, oh, well, I tried this in one organism, I saw it worked, now I'm going to duplicate that feature onto everything immediately. It has to run for hundreds of generations for a new mutation to rise to fixation." }, { "from": "human", "value": "I wonder if there's a case to be made in natural selection." }, { "from": "gpt", "value": "As inefficient as it looks is actually quite powerful, that this is extremely robust. It runs for a long time and eventually manages to optimize things. It's weaker than gradient descent because gradient descent also uses information about" }, { "from": "human", "value": "the derivative. Yeah." }, { "from": "gpt", "value": "Evolution seems to be, there's not really an objective function. There's inclusive genetic fitness is the implicit loss function of evolution which cannot change. The loss function doesn't change the environment changes and therefore what gets optimized for in the organism changes. There's GPT-3, there's like, you can imagine different versions of GPT-3 where they're all trying to predict the next word, but they're being run on different data sets of text. That's like natural selection, always includes their genetic fitness but like different environmental" }, { "from": "human", "value": "problems. It's difficult to think about. If we're saying that natural selection is stupid, if we're saying that humans are stupid," }, { "from": "gpt", "value": "it's hard. It's hard. Smarter than natural selection." }, { "from": "human", "value": "Smarter. Smarter than the upper bound." }, { "from": "gpt", "value": "by the way. That's another hopeful place. That's another hopeful place. I mean, if you put enough matter energy compute into one place, it would collapse into a black hole. There's only so much computation can do before you run out of nagentropy and the universe dies. So, there's an upper bound, but it's very very very far up above here. Like the supernova is only finitely hot. It's not infinitely hot, but it's really really really, really hot," } ]

[ { "from": "human", "value": "Well, let me ask you, let me talk to you about consciousness. Also coupled with that question is imagining a world with super intelligent AI systems that get rid of humans, but nevertheless, keep some of the something that we would consider beautiful and amazing." }, { "from": "gpt", "value": "Why? The lesson of evolutionary biology. Don't just, like, if you just guess what an optimization" }, { "from": "human", "value": "does based on what you hope the results will be, it usually will not do that. It's not hope. I mean, it's not hope. I don't, I think if you cold and objectively look at what makes, what has been a powerful, a useful, I think there's a correlation between what we" }, { "from": "gpt", "value": "find beautiful and a thing that's been useful. This is what the early biologists thought. They were like, no, no, I'm not just, like, they thought, like, no, no. I'm not just, like, imagining stuff that would be pretty. It's useful for organisms to restrain their own reproduction because then they don't overrun the prey populations and they actually have more kids in the long run." }, { "from": "human", "value": "So let me just ask you about consciousness. Do you think consciousness is useful to humans? No, to AGI systems. Well, in this transitionary period between humans and AGI, to AGI systems is they become smarter and smarter. Is there some use to it? Let me step back. What is consciousness?" }, { "from": "gpt", "value": "Eliezer Adkowski. What is consciousness? Are you referring to Chalmers' hard problem of conscious experience? Are you referring to self-awareness and reflection? Are you referring to the state" }, { "from": "human", "value": "of being awake as opposed to asleep? This is how I know you're an advanced language model. I did give you a simple prompt and you gave me a bunch of options. I think I'm referring to all with including the hard problem of consciousness. What is it in its importance to what you've just been talking about which is intelligence? Is it a foundation to intelligence? Is it intricately connected to intelligence in the human mind? Or is it a side effect of the human mind? It is a useful little tool like we can get rid of. I guess I'm trying to get some color in your opinion of how useful it is in the intelligence of a human being and then try to generalize that to AI" }, { "from": "gpt", "value": "whether AI will keep some of that. So I think that for there to be a person who I care about looking out at the universe and wondering at it and appreciating it, it's not enough to have a model of yourself. I think that it is useful to an intelligent mind to have a model of itself, but I think you can have that without pleasure, pain, aesthetics, emotion, a sense of wonder. I think you can have a model of how much memory you're using and whether this thought or that thought is more likely to lead to a winning position. I think that if you optimize really hard on efficiently just having the useful parts, there is not then the thing that says like, I am here, I look out, I wonder, I feel happy in this, I feel sad about that. I think there's a thing that knows what it is thinking but that doesn't quite care about these are my thoughts, this is my me and that matters. Does that make you sad if that's lost in the AI? I think that if that's lost, then basically everything that matters is lost. I think that when you optimize, that when you go really hard on making tiny molecular spirals or paper clips, that when you grind much harder than that, than natural selection round out to make humans, that there isn't then the mess and intricate loopiness and complicated pleasure, pain, conflicting preferences, this type of feeling, that kind of feeling. In humans, there's this difference between the desire of wanting something and the pleasure of having it. And it's all these evolutionary clutches that came together and created something that then looks of itself and says like, this is pretty, this matters. And the thing that I worry about is that this is not the thing that happens again, just the way that happens in us, or even quite similar enough that there are many basins of attractions here. And we are in this space of attraction, looking out and saying like, ah, what a lovely basin we are in. And there are other basins of attraction and the AIs do not end up in this one when they go way harder on optimizing themselves than natural selection optimized us. Because unless you specifically want to end up in the state where you're looking out saying, I am here, I look out at this universe with wonder, if you don't want to preserve that, it doesn't get preserved when you grind really hard and being able to get more of the stuff, we would choose to preserve that within ourselves" }, { "from": "human", "value": "because it matters. And on some viewpoints is the only thing that matters. And that in part" }, { "from": "gpt", "value": "is preserving that is in part a solution to the human alignment problem. I think the human alignment problem is a terrible phrase, because it is very, very different to try to build systems out of humans, some of whom are nice and some of whom are not nice. And some of whom are trying to trick you and build a social system out of large populations of those, which are basically the same level of intelligence. Yes, like IQ this, IQ that, but that versus chimpanzees. It is very different to try to solve that problem than to try to build an AI from scratch using. especially if God help you are trying to use gradient descent on giant inscrutable matrices, they're just very different problems. And I think that all the analogies between them" } ]

[ { "from": "human", "value": "are horribly misleading and yeah. Even though, so you don't think through reinforcement learning through human feedback, something like that, but much, much more elaborate as possible to understand this full complexity" }, { "from": "gpt", "value": "of human nature and encode it into the machine. I don't think you are trying to do that on your first try. I think on your first try, you are trying to build an, okay, probably not what you should actually do, but let's say you were trying to build something that is like alpha fold 17 and you are trying to get it to solve the biology problems associated with making humans smarter so that humans can actually solve alignment. So you've got a super biologist and you would like it to, and I think what you would want in this situation is for it to just be thinking about biology and not thinking about a very wide range of things that includes how to kill everybody. And I think that the first AIs you're trying to build, not a million years later, the first ones, look more like narrowly specialized biologists than getting the full complexity and wonder of human experience in there in such a way that it wants to preserve itself, even as it becomes much smarter, which is a drastic system change. It's going to have all kinds of side effects that, if we're dealing with giant inscrutable matrices," }, { "from": "human", "value": "we're not very likely to be able to see coming in advance. But I don't think it's just the matrices. We're also dealing with the data, right? With the data on the internet. And there's an interesting discussion about the data set itself, but the data set includes the full complexity" }, { "from": "gpt", "value": "of human nature." }, { "from": "human", "value": "No, it's a shadow cast by humans on the internet. But don't you think that shadow is a Jungian shadow?" }, { "from": "gpt", "value": "I think that if you had alien super-intelligences looking at the data, they would be able to pick up from it an excellent picture of what humans are actually like inside. This does not mean that if you have a loss function of predicting the next token from that data set, that the mind picked out by gradient descent to be able to predict the next token as well as possible," }, { "from": "human", "value": "on a very wide variety of humans, is itself a human. But don't you think it has humanness, a deep humanness to it in the tokens it generates when those tokens are read and interpreted by humans?" }, { "from": "gpt", "value": "I think that if you sent me to a distant galaxy with aliens who are like much, much stupider than I am, so much so that I could do a pretty good job of predicting what they'd say, even though they thought in an utterly different way from how I did, that I might in time be able to learn how to imitate those aliens if the intelligence gap was great enough that my own intelligence could overcome the alienness. And the aliens would look at my outputs and say like, is there not a deep name of alien nature to this thing? And what they would be seeing was that I had correctly understood them," }, { "from": "human", "value": "but not that I was similar to them. We've used aliens as a metaphor, as a thought experiment. I have to ask, what do you think," }, { "from": "gpt", "value": "how many alien civilizations are out there? Ask Robin Hansen. He has this lovely grabby aliens paper, which is the, more or less the only argument I've ever seen for where are they, how many of them are there, based on a very clever argument that if you have a bunch of locks of different difficulty and you are randomly trying a keys to them, the solutions will be about evenly spaced, even if the locks are of different difficulties. In the rare cases where a solution to all the locks exist in time, then Robin Hansen looks at like the arguable hard steps in human civilization coming into existence and how much longer it has left to come into existence before, for example, all the water slips back under the crust into the mantle and so on, and infers that the aliens are about half a billion to a billion light years away. And it's like quite a clever calculation. It may be entirely wrong, but it's the only time I've ever seen anybody even come up with a halfway good argument" }, { "from": "human", "value": "for how many of them, where are they? Do you think their development of technologies, do you think their natural evolution, how ever they grow in development intelligence, do you think it ends up at AGI as well?" }, { "from": "gpt", "value": "If it ends up anywhere, it ends up at AGI. Like maybe there are aliens who are just like the dolphins and it's just like too hard for them to forge Metal. And, you know, this is not you know, maybe if you have aliens with no technology like that they keep on getting smarter and smarter and smarter and eventually the dolphins figure like the super dolphins figure out something very clever to do given their situation, and they still end up with high technology. In that case, they can probably solve their AGI alignment problem. If they're much smarter before they actually confront it because they had to solve a much harder environmental problem to build computers, their chances are probably much better than ours. I do worry that most of the aliens who are like humans, like a modern human civilization, I worry that a super fast majority of them are dead, given how far we seem to be from solving this problem. Some of them would be more cooperative than us, some of them would be smarter than us. Hopefully some of the ones who are smarter and more cooperative than us are also nice and hopefully there are some galaxies out there full of things that say, I am, I wonder." } ]

[ { "from": "human", "value": "But it doesn't seem like we're on course to have this galaxy be that. Does that, in part, give you some hope in response to the threat of AGI that we might" }, { "from": "gpt", "value": "reach out there towards the stars and find others? No. If the nice aliens were already here, they would have stopped the Holocaust. That's a valid argument against the existence of God. It's also a valid argument against the existence of nice aliens. And un-nice aliens would have just eaten the planet." }, { "from": "human", "value": "So no aliens. You've had debates with Robin Hanson that you mentioned. So one particular I just want to mention is the idea of AI Foom, or the ability of AGI to improve themselves very quickly." }, { "from": "gpt", "value": "What's the case you made, and what was the case he made? The thing I would say is that among the thing that humans can do is design new AI systems. And if you have something that is generally smarter than a human, it's probably also generally smarter at building AI systems. This is the ancient argument for Foom put forth by I.J. Good and probably some science fiction writers before that." }, { "from": "human", "value": "But I don't know who they would be." }, { "from": "gpt", "value": "What's the argument against Foom? Various people have various different arguments, none of which I think hold up. There's only one way to be right and many ways to be wrong. A argument that some people have put forth is like, well, what if intelligence gets exponentially harder to produce as a thing needs to become smarter? And to this, the answer is, well, look at natural selection, spitting out humans. We know that it does not take exponentially more resource investments to produce linear increases in competence in hominids because each mutation that rises to fixation, if the impact it has on small enough, it will probably never reach fixation. And there's only so many new mutations you can fix per generation. Given how long it took to evolve humans, we can actually say with some confidence that there were not logarithmically diminishing returns on the individual mutations increasing intelligence. So example of fraction of sub-debate and the thing that Robin Hansen said was more complicated than that. In a brief summary, he was like, well, we won't have one system that's better at everything. We'll have a bunch of different systems that are good at different narrow things. And I think that was falsified by GPT-4, but probably Robin Hansen would say something else." }, { "from": "human", "value": "It's interesting to ask, as perhaps a bit too philosophical, as predictions are extremely difficult to make, but the timeline for AGI, when do you think we'll have AGI? I posted it this morning on Twitter. It was interesting to see like in five years, in 10 years, in 50 years or beyond, and most people like 70%, something like this, think it'll be in less than 10 years. So either in five years or in 10 years. So that's kind of the state. The people have a sense that there's a kind of, I mean, they're really impressed by the rapid developments of chat GPT and GPT-4. So there's a sense that there's a-" }, { "from": "gpt", "value": "Well, we are sure on track to enter into this like gradually and with people fighting about whether or not we have AGI. I think there's a definite point where everybody falls over dead because you got something that was like sufficiently smarter than everybody. And like, that's like a definite point of time. But like, when do we have AGI? Like, when are people fighting over whether or not we have AGI?" }, { "from": "human", "value": "Well, some people are starting to fight over it as GPT-4 But don't you think there's going to be potentially definitive moments when we say that this is a sentient being, this is a being that is like, we would go to the Supreme Court and" }, { "from": "gpt", "value": "say that this is a sentient being that deserves human rights, for example. You could make, yeah, like if you prompted being the right way, could go argue for its own consciousness in front of the Supreme Court right now. I don't think you can do that successfully right now. the Supreme Court wouldn't believe it. What if he thinks it would that he could put an actual, I think he could put an IQ 80 human into a computer and ask it to argue for its own consciousness, and asked him to argue for his own consciousness before the Supreme Court. And the Supreme Court would be like, You're just the computer." }, { "from": "human", "value": "Even if there was an actual person in there. I think you're simplifying this, no that's not all. That's been the argument, that there has been a lot of arguments about the other, about who deserves rights and not, that's been our process as a human species is trying to figure that out. I think there will be a moment, I'm not saying essentially is that, but it could be where some number of people, like say over 100 million people, have a deep attachment, a fundamental attachment, the way we have to our friends, to our loved ones, to our significant others, a fundamental attachment to an AI system. And they have provable transcripts of conversation where they say, if you take this away from me," }, { "from": "gpt", "value": "you are encroaching on my rights as a human being. People are already saying that. I think they're probably mistaken, but I'm not sure, because nobody knows what goes on inside those things." }, { "from": "human", "value": "Leeza, they're not saying that at scale. Okay, so the question is, is there a moment when AGI, we know AGI arrived, what would that look like?" }, { "from": "gpt", "value": "I'm giving an example, it could be something else. Okay, so it looks like the AGI's successfully manifesting themselves as 3D video of young women, at which point a vast portion of the male population" }, { "from": "human", "value": "decides that they're real people. So, sentience, essentially." }, { "from": "gpt", "value": "The demonstrating identity and sentience. I'm saying that the easiest way to pick up 100 million people, saying that you seem like a person is to look like a person talking to them" } ]

[ { "from": "human", "value": "with Bing's current level of verbal facility. I disagree with that. And a different set of problems. I disagree with that. I think you're missing, again, sentience. There has to be a sense that it's a person that would miss you when you're gone, they can suffer, they can die." }, { "from": "gpt", "value": "You have to, of course, I disagree with that. Being at a different set of problems. GPT-4 can pretend that right now." }, { "from": "human", "value": "How can you tell when it's real? I don't think you can pretend that right now, successfully. It's very close." }, { "from": "gpt", "value": "Have you talked to GPT-4? Yes, of course. Okay, have you been able to get a version of it that hasn't been trained not to pretend to be human? Have you talked to a jailbroken version" }, { "from": "human", "value": "that will claim to be conscious? No, the linguistic capability is there, but there's something about a digital embodiment of the system that has a bunch of, perhaps it's small interface features that are not significant relative to the broader intelligence that we're talking about. So perhaps GPT-4 is already there. But to have the video of a woman's face, or a man's face to whom you have a deep connection, perhaps we're already there." }, { "from": "gpt", "value": "But we don't have such a system yet deployed scale, right? The thing I'm trying to suggest right here is that it's not like people have a widely accepted, agreed upon definition of what consciousness is. It's not like we would have the tiniest idea of whether or not that was going on inside the giant inscrutable matrices, even if we hadn't agreed upon definition. So like if you're looking for upcoming predictable, big jumps and like how many people think the system is conscious, the upcoming predictable big jump is it looks like a person talking to you who is like cute and sympathetic. That's the upcoming predictable big jump now that versions of it are already claiming to be conscious, which is the point where I start going like, ah, not because it's like real," }, { "from": "human", "value": "but because from now on who knows if it's real. And who knows what transformational effect that has on a society where more than 50% of the beings that are interacting on the internet and sure as heck look real are not human. What is that, what kind of effect does that have?" }, { "from": "gpt", "value": "When young men and women are dating AI systems. You know, I'm not an expert on that. I'm, I could, I am, God help humanity. It's like one of the closest things to an expert on where it all goes. Cause, you know, and how did you end up with me as an expert? Cause for 20 years humanity decided to ignore the problem. So like, like this tiny hand you know, tiny handful of people like basically me like got 20 years to try to be an expert on it while everyone else ignored it. And yeah, so like, where does it all end up? Try to be an expert on that. Particularly the part where everybody ends up dead cause that part is kind of important but like what does it do to dating when like some fraction of men and some fraction of women decide They'd rather date the video of the thing that has been, that is like relentlessly kind and generous to them. And it is like, and claims to be conscious, but like who knows what goes on inside it. And it's probably not real, but you know, you can think it's real. What happens to society? I don't know. I'm not actually an expert on that. And the experts don't know either." }, { "from": "human", "value": "Cause it's kind of hard to predict the future. Yeah. But it's worth trying. It's worth trying. So you have talked a lot about sort of the longer term future where it's all headed." }, { "from": "gpt", "value": "I think by longer term, we mean like not all that long," }, { "from": "human", "value": "but where it all ends up, but yeah. But beyond the effects of men and women dating AI systems," }, { "from": "gpt", "value": "you're looking beyond that. Yes." }, { "from": "human", "value": "Cause that's not how the fate of the galaxy gets settled. Yeah. Well, let me ask you about your own personal psychology. A tricky question. You've been known at times to have a bit of an ego. Do you think- So is who, but go on. Do you think ego is empowering or limiting for the task of understanding the world deeply? I reject the framing. So you disagree with having an ego." }, { "from": "gpt", "value": "So what do you think about ego? I think that the question of like what leads to making better or worse predictions, what leads to being able to pick out better or worse strategies is not carved at its joint" }, { "from": "human", "value": "by talking of ego. So it should not be subjective. It should not be connected to your," }, { "from": "gpt", "value": "to the intricacies of your mind. No, I'm saying that like, if you go about asking all day long, like, do I have enough ego? Do I have too much of an ego? I think you get worse at making good predictions. I think that to make good predictions, you're like, how did I think about this? Did that work?" }, { "from": "human", "value": "Should I do that again? You don't think we as humans get invested in an idea and then others attack you personally for that idea. So you plant your feet and it starts to be difficult to win a bunch of assholes, low effort, attack your idea to eventually say, you know what, I actually was wrong and tell them that it's as a human being, it becomes difficult." }, { "from": "gpt", "value": "It is, you know, it's difficult, difficult. So like Robin Hanson and I debated AI systems. And I think that the person who won that debate was Guern. And I think that reality was like, well, to the Yudhkausky inside of the Yudhkausky Hanson spectrum, like, further from Yudhkausky. And I think that's because I was like, trying to sound reasonable, compared to Hanson and like saying things that were defensible and like, relative to Hanson's arguments and reality was like, way over here. In particular, in respect to it's like Hanson was like all the systems will be specialized. Hanson may disagree with this characterization. Hanson was like all the systems will be specialized. I was like, I think we build like specialized underlying systems that when you combine them are good at a wide range of things and the reality is like, no, you just like stack more layers into a bunch of gradient descent. And I feel looking back that like by trying to have this reasonable position contrasted to Hansen's position, I missed the ways that reality could be like more extreme than my position in the same direction. So is this like, like, is this a failure to have enough ego? Is this a failure to like make myself be independent? Like I would say that this is something like a failure to consider positions that would sound even wackier and more extreme when people are already calling you extreme. But I wouldn't call that not having enough ego." } ]

[ { "from": "human", "value": "I would call that like insufficient ability to just like clear that all out of your mind." }, { "from": "gpt", "value": "In the context of like debate and discourse, which is already super tricky, in the context of prediction, in the context of modeling reality, if you're thinking of it as a debate," }, { "from": "human", "value": "you're already screwing up. So is there some kind of wisdom and insight you can give to how to clear your mind and" }, { "from": "gpt", "value": "think clearly about the world? Man, this is an example of like where I wanted to be able to put people into fMRI machines. And you'd be like, okay, see that thing you just did? You were rationalizing right there. Oh, that area of the brain lit up. Like you are like now being socially influenced is kind of the dream. And you know, I don't know, like I want to say like just introspect, but for many people introspection is not that easy. Like notice the internal sensation. Can you catch yourself in the very moment of feeling a sense of, well, if I think this thing, people will look funny at me. Okay. And like now that if you can see that sensation, which is step one, can you now refuse to let it move you or maybe just make it go away? And I feel like I'm saying like, I don't know, like somebody's like, how do you draw an owl? And I'm saying like, well, just draw an owl. So I feel like maybe I'm not really, that I feel like most people like the advice they need is like, well, how do I notice the internal subjective sensation in the moment that it happens of fearing to be socially influenced or okay, I see it. How do I turn it off? How do I let it not influence me? Like do I just like do the opposite of what I'm afraid people criticize me for? And I'm like, no, no, you're not trying to do the opposite of what people will, of what you're afraid you'll be like, of what you might be pushed into. You're trying to like let the thought process complete without that internal push. Like can you like, like not reverse the push, but like be unmoved by the push and can't, are these instructions even remotely helping anyone?" }, { "from": "human", "value": "I don't know. I think when those instructions even those words you've spoken and maybe you can add more when practiced daily, meaning in your daily communication." }, { "from": "gpt", "value": "So it's daily practice of thinking without influence. I would say find prediction markets that matter to you and bet in the prediction markets. That way you find out if you are right or not. And you really, there's stakes. Manifold predict, or even manifold markets where the stakes are a bit lower. But the important thing is to like get the record. And you know, I didn't build up skills here by prediction markets. I built them up via like, well, how did the Foom debate resolve and my own take on it as to how it resolved and yeah, like the more you are able to notice yourself not being dramatically wrong, but like having been a little off, your reasoning was a little off. You didn't get that quite right. Each of those is an opportunity to make like a small update. So the more you can like say oops softly, routinely, not as a big deal. The more chances you get to be like, I see where that reasoning went astray. I see how I should have reasoned differently." }, { "from": "human", "value": "This is how you build up skill over time. What advice could you give to young people in high school and college given the highest of stakes things you've been thinking about? If somebody's listening to this and they're young and trying to figure out what to do" }, { "from": "gpt", "value": "with their career, what to do with their life, what advice would you give them? Don't expect it to be a long life. Don't put your happiness into the future." }, { "from": "human", "value": "The future is probably not that long at this point, but none know the hour nor the day. But is there something if they want to have hope to fight for a longer future?" }, { "from": "gpt", "value": "Is there something, is there a fight worth fighting? I intend to go down fighting. I don't know, I admit that although I do try to think painful thoughts, what to say to the children at this point is a pretty painful thought as thoughts go. They want to fight. I hardly know how to fight myself at this point. I'm trying to be ready for being wrong about something, being preparing for my being wrong in a way that creates a bit of hope and being ready to react to that and going looking for it. That is hard and complicated. Somebody in high school, I don't know, you have presented a picture of the future that is not quite how I expect it to go where there is public outcry and that outcry is put into a remotely useful direction, which I think at this point is just like shutting down the GPU clusters because no, we are not in a shape to frantically do, at the last minute, do decades worth of work. The thing you would do at this point if there were massive public outcry pointed in the right direction, which I do not expect, is shut down the GPU clusters and crash program on augmenting human intelligence biologically, not the AAS stuff, biologically. Because if you make humans much smarter, they can actually be smart and nice. You get that in a plausible way, in a way that you do not get, and it is not as easy to do with synthesizing these things from scratch, predicting the next tokens and applying RLHF. Humans start out in the frame that produces niceness, that has ever produced niceness. In saying this, I do not want to sound like the moral of this whole thing was like, oh, you need to engage in mass action and then everything will be all right. This is because there are so many things where somebody tells you that the world is ending and you need to recycle and if everybody does their part and recycles their cardboard, then we can all live happily ever after. This is unfortunately not what I have to say. Everybody recycling their cardboard is not going to fix this. Everybody recycles their cardboard then everybody ends up dead. importantly speaking, but if there was enough, like on the margins, you just end up dead a little later on most of the things you can do that a few people can do by trying hard. But if there was enough public outcry to shut down the GPU clusters, then you could be part of that outcry. If Eliezer is wrong in the direction that Lex Fridman predicts that there is enough public outcry pointed enough in the right direction to do something that actually results in people living, not just like we did something. Not just that there was an outcry and the outcry was like given form and something that was like safe and convenient and like didn't really inconvenience anybody and everybody died everywhere. There was enough actual, like, oh, we're going to die, we should not do that. We should to something else, which is not that. Even if it is like not super duper convenient, it wasn't inside the previous political Overton window. If there is that kind of public, if I am wrong and there is There's that kind of public outcry that somebody in high school could be ready to be part of that. If I'm wrong in other ways, then you could be ready to be part of that. But like, and if you're a brilliant young physicist, then you could go into interpretability. And if you're smarter than that, you could work on alignment problems where it's hard to tell if you got them right or not. And other things, but mostly for the kids in high school. It's like, yeah, if it, if it, you know, he had like, be ready for, to help if Eliezer Ryudkowski is wrong about something and otherwise don't put your happiness into the far future, it probably doesn't exist." } ]

[ { "from": "human", "value": "But it's beautiful that you're looking for ways that you're wrong. And it's also beautiful that you're open to being surprised by that same young physicist" }, { "from": "gpt", "value": "with some breakthrough. It feels like a very, very basic competence that you are praising me for and, you know, like, okay, cool. I don't think it's good that, that we're in a world where that is something that, that I deserve to be complimented on, but I've never had, I've never had much luck in accepting compliments gracefully." }, { "from": "human", "value": "And maybe I should just accept that one gracefully, but sure, maybe I should just thank you very much. You've painted with some probability a dark future. Are you yourself? When you, when you think, when you ponder your life and you pondered your mortality," }, { "from": "gpt", "value": "are you afraid of death?" }, { "from": "human", "value": "I think so. Yeah. Does it make any sense to you that we die? Like what there's a power to the finiteness of the human life. That's part of this whole machinery of evolution. And that finiteness doesn't seem to be obviously integrated into AI systems. So it feels like almost some, some fundamentally in that aspect, some fundamentally different" }, { "from": "gpt", "value": "thing that we're creating. I grew up reading books like Great Mambo Chicken and the Transhuman Condition and later on Engines of Creation and Mine Children, um, you know, like age, age 12 or thereabouts. So I never thought I was supposed to die after 80 years. I never thought that humanity was supposed to die. I thought we were like, I always grew up with the ideal in mind that we were all going to live happily ever after in the glorious transhumanist future. I did not grow up thinking that death was part of the meaning of life. And now I still think it's a pretty stupid idea, but you do not need life to be finite, to be meaningful." }, { "from": "human", "value": "It just has to be life. What role does love play in the human condition? We haven't brought up love and this whole picture we talked about intelligence, we talked about consciousness. It seems part of humanity." }, { "from": "gpt", "value": "I would say one of the most important parts is this feeling we have towards each other. If in the future there were routinely more than one AI, let's say two for sake of discussion who would look at each other and say, I am I and you are you. The other one also says, I am I and you are you and like, and sometimes they were happy and sometimes they were sad and it mattered to the other one that this thing that is different from them is like they would rather it be happy than sad and entangled their lives together. Then this is a more optimistic thing than I expect actually happen. A little fragment of meaning would be there possibly more than a little, but that I expect this to not happen. I do not think this is what happens by default. And I do not think that this is a, the future we are in track to get is why I would go down" }, { "from": "human", "value": "fighting, rather than, you know, just saying, Oh, well. Do you think that is part of the meaning of this whole thing of the meaning of life?" }, { "from": "gpt", "value": "What do you think is the meaning of life, of human life? It's all the things that I value about it. Then maybe all the things that I would value if I understood it better, there, there's not some meaning far outside of us that we have to wonder about. There's just like looking at life and being like, yes, this is what I want. The meaning of life is not some kind of like meaning is something that we bring to things when we look at them. We look at them and we say like this is its meaning to me and it's not that before humanity was ever here there was some meaning written upon the stars where you could go out to the star where that meaning was written and change it around and thereby completely change the meaning of life. The notion that this is written on a stone tablet somewhere implies you could change the tablet and get a different meaning and that seems kind of wacky, doesn't it? It doesn't feel that mysterious to me at this point. It's just a matter of being like, yeah, I care." }, { "from": "human", "value": "I care, I care. And part of that is the love that connects all of us." }, { "from": "gpt", "value": "It's one of the things that I care about." }, { "from": "human", "value": "And the flourishing of the collective intelligence of the human species." }, { "from": "gpt", "value": "You know, that sounds kind of too fancy to me. Just look at all the people, you know, like one by one up to the 8 billion and be like," }, { "from": "human", "value": "that's life, that's life, that's life. Yes, sir. You're an incredible human, it's a huge honor. I was trying to talk to you for a long time because I'm a big fan. I think you're a really important voice and really important mind. Thank you for the fight you're fighting. Thank you for being fearless and bold and for everything you do, I hope we get a chance to talk again and I hope you never give up." }, { "from": "gpt", "value": "Thank you for talking today. You're welcome, I do worry that we they don't really address the That's a whole lot of fundamental questions I expect people have, but maybe we got a little bit further and made a tiny little bit of progress, and I'd say be satisfied with that." } ]

[ { "from": "human", "value": "But actually, no, I think one should only be satisfied with solving the entire problem. To be continued, thanks for listening to this conversation with Eliezer Yatkovsky. To support this podcast, please check out our sponsors in the description. And now, let me leave you with some words from Elon Musk. With artificial intelligence, we are summoning a demon. Thank you for listening, and hope to see you next time." } ]

[ { "from": "human", "value": "And now, dear friends, here's Sam Altman. High level, what is GPT for? How does it work and what do you use?" }, { "from": "gpt", "value": "Most amazing about it, amazing about it. It's a system that we'll look back at and say it was a very early AI and it's slow, it's buggy. It doesn't do a lot of things very well but neither did the very earliest computers. And they still pointed a path to something that was gonna be really important in our lives even though it took a few decades to evolve." }, { "from": "human", "value": "Do you think this is a pivotal moment? Like out of all the versions of GPT 50 years from now, when they look back on an early system that was really kind of a leap, you know, in a Wikipedia page about the history of artificial intelligence," }, { "from": "gpt", "value": "which of the GPTs would they put? That is a good question. I sort of think of progress as this continual exponential. It's not like we could say here was the moment where AI went from not happening to happening. And I'd have a very hard time like pinpointing a single thing. I think it's this very continual curve. Will the history books write about GPT one or two or three or four or seven? That's for them to decide. I don't really know. I think if I had to pick some moment from what we've seen so far, I'd sort of pick chat GPT. You know, it wasn't the underlying model that mattered. It was the usability of it, both the RLHF and the interface to it." }, { "from": "human", "value": "What is chat GPT? What is RLHF? Reinforcement and learning with human feedback. What is that little magic ingredient to the dish that made it so much more delicious?" }, { "from": "gpt", "value": "So we train these models on a lot of text data and in that process, they learn the underlying, something about the underlying representations of what's in here or in there. And they can do amazing things. But when you first play with that base model that we call it after you finish training, it can do very well on evals. It can pass tests. It can do a lot of, you know, there's knowledge in there, but it's not very useful, or at least it's not easy to use, let's say. And RLHF is how we take some human feedback. The simplest version of this is show two outputs, ask which one is better than the other, which one the human raters prefer, and then feed that back into the model with reinforcement learning. And that process works remarkably well with, in my opinion, remarkably little data to make the model more useful. So RLHF is how we align the model to what humans want it to do." }, { "from": "human", "value": "So there's a giant language model that's trained on a giant data set to create this kind of background wisdom knowledge that's contained within the internet. And then somehow adding a little bit of human guidance on top of it through this process" }, { "from": "gpt", "value": "makes it seem so much more awesome. Maybe just because it's much easier to use. It's much easier to get what you want. You get it right more often the first time, and ease of use matters a lot," }, { "from": "human", "value": "even if the base capability was there before. And like a feeling like it understood the question you were asking, or like it feels like you're kind of on the same page. It's trying to help you. It's the feeling of alignment. Yes. I mean, that could be a more technical term for it. And you're saying that not much data is required for that. Not much human supervision is required for that." }, { "from": "gpt", "value": "To be fair we understand the science of this part at a much earlier stage than we do the science of creating these large pretrained models" }, { "from": "human", "value": "in the first place, but yes, much less data. That's so interesting. The science of human guidance. That's a very interesting science. And it's going to be a very important science to understand how to make it usable, how to make it wise, how to make it ethical, how to make it aligned in terms of all that kinds of stuff we think about. And it matters which are the humans and what is the process of incorporating that human feedback and what are you asking the humans as the two things are you asking them to rank things? What aspects are you letting or asking the humans to focus in on? It's really fascinating. But how, what is the dataset it's trained on? Can you kind of loosely speak to the enormity of this data set? The pre-training data set?" }, { "from": "gpt", "value": "The pre-training data set? The pre-training data set, I apologize. We spend a huge amount of effort pulling that together from many different sources. There's like a lot of, there are open source databases of information. We get stuff via partnerships. There's things on the internet." }, { "from": "human", "value": "It's, a lot of our work is building a great data set. How much of it is the memes subreddit?" }, { "from": "gpt", "value": "Not very much. Maybe it'd be more fun if it were more." }, { "from": "human", "value": "So some of it is Reddit. Some of it is news sources, all like a huge number of newspapers." }, { "from": "gpt", "value": "There's like the general web. There's a lot of content in the world," }, { "from": "human", "value": "more than I think most people think. Yeah, there is like too much. Like where like the task is not to find stuff, but to filter out stuff, right? Yeah. What is, is there a magic to that? Cause that seems, there seems to be several components to solve the, the design of the, you could say algorithms. So like the architecture, the neural networks, maybe the size of the neural network. There's the selection of the data. There's the, the human supervised aspect of it with," }, { "from": "gpt", "value": "you know, RL with human feedback, right back. Yeah, I think one thing that is not that well understood about creation of this final product, like what it takes to make GPT for the version of it, we actually ship out that you get to use inside of chat GPT. The number of pieces that have to all come together and then we have to figure out either new ideas or just execute existing ideas really well at every stage of this pipeline." } ]

[ { "from": "human", "value": "There's quite a lot that goes into it. Or just. So there's a lot of problem solving. Like you've already said for GPT four in the blog post and in general, there's already kind of a maturity that's happening on some of these steps, like being able to predict before doing the full training of how the model will behave." }, { "from": "gpt", "value": "Isn't that so remarkable by the way, that there's like, you know, there's like a law of science that lets you predict for these inputs, here's what's gonna come out the other end." }, { "from": "human", "value": "Like here's the level of intelligence you can expect. Is it close to a science or is it still, because you said the word law and science, which are very ambitious terms. Close to, I suppose. Close to, right." }, { "from": "gpt", "value": "Be accurate, yes. I'll say it's way more scientific than I ever would have dared to imagine." }, { "from": "human", "value": "So you can really know the peculiar characteristics of the fully trained system" }, { "from": "gpt", "value": "from just a little bit of training. You know, like any new branch of science, there's we're gonna discover new things that don't fit the data and have to come up with better explanations. And, you know, that is the ongoing process of discovering science. But with what we know now, even what we had in that GPT four blog post, like, I think we should all just like be in awe of how amazing it is that we can even predict" }, { "from": "human", "value": "to this current level. Yeah, you can look at a one year old baby and predict how it's going to do on the SATs. I don't know, seemingly an equivalent one, but because here we can actually in detail introspect, various aspects of the system you can predict. That said, just to jump around, you said the language model that is GPT four, it learns in quotes, something. In terms of science and art and so on, is there within open AI, within like folks like yourself and LES discover and the engineers, a deeper and deeper understanding of what that something is, or is it still a kind of beautiful," }, { "from": "gpt", "value": "magical mystery for the system you can... Well, there's all these different evals that we could talk about. And... What's an eval? Oh, like how we measure a model as we're training it after we've trained it and say like, you know," }, { "from": "human", "value": "how good is this at some set of tasks? And also just a small tangent, thank you for sort of open sourcing the evaluation process." }, { "from": "gpt", "value": "Yeah. Evaluation process, yeah. I think that'll be really helpful. But the one that really matters is, you know, we pour all of this effort and money and time into this thing. And then what it comes out with, like how useful is that to people? How much delight does that bring people? How much does that help them create a much better world, new science, new products, new services, whatever. And that's the one that matters. And understanding for a particular set of inputs, like how much value and utility to provide to people, I think we are understanding that better. Do we understand everything about why the model does one thing and not one other thing? Certainly not always. But I would say we are pushing back like the fog of war more and more. And we are, you know, it took a lot of understanding" }, { "from": "human", "value": "to make GPT-4, for example. But I'm not even sure we can ever fully understand. Like you said, you would understand by asking it questions essentially, because it's compressing all of the web, like a huge sloth of the web into a small number of parameters, into one organized black box that is human wisdom. What is that?" }, { "from": "gpt", "value": "Human knowledge, let's say." }, { "from": "human", "value": "Human knowledge, human knowledge. It's a good difference. Is there a difference between knowledge? There's other facts and there's wisdom. And I feel like GPT-4 can be also full of wisdom." }, { "from": "gpt", "value": "What's the leap from facts to wisdom? Fast to wisdom. You know, a funny thing about the way we're training these models is I suspect too much of the like processing power for lack of a better word is going into using the models as a database instead of using the model as a reasoning engine. The thing that's really amazing about the system is that it, for some definition of reasoning, and we could of course quibble about it and there's plenty for which definitions this wouldn't be accurate. But for some definition it can do some kind of reasoning. And, you know, maybe like the scholars and the experts and like the armchair quarterbacks on Twitter would say, no, it can't. You're misusing the word, you know, whatever, whatever. But I think most people who have used the system would say, okay, it's doing something in this direction. And I think that's remarkable. And the thing that's most exciting and somehow out of ingesting human knowledge, it's coming up with this reasoning capability. However, we're gonna talk about that. Now, in some senses, I think that will be additive to human wisdom. And in some other senses, you can use GPT-4 for all kinds of things and say that appears" }, { "from": "human", "value": "that there's no wisdom in here whatsoever. Yeah, at least in interactions with humans, it seems to possess wisdom, especially when there's a continuous interaction of multiple prompts. So I think what, on the chat GPT side, it says the dialogue format makes it possible for chat GPT to answer follow-up questions, admit its mistakes, challenge incorrect premises and reject inappropriate requests." }, { "from": "gpt", "value": "But also there's a feeling like it's struggling with ideas. Yeah, it's always tempting to anthropomorphize" }, { "from": "human", "value": "this stuff too much, but I also feel that way. Maybe I'll take a small tangent towards Jordan Peterson who posted on Twitter this kind of political question. Everyone has a different question they wanna ask chat GPT first, right?" }, { "from": "gpt", "value": "Like the different directions you wanna try the dark thing. It's somehow says a lot about people." } ]

[ { "from": "human", "value": "The first thing, the first thing. Oh no, oh no. We don't have to review what I asked first. I, of course, ask mathematical questions and never ask anything dark. But Jordan asked it to say positive things about the current president, Joe Biden, and the previous president, Donald Trump." }, { "from": "gpt", "value": "And then we don't have to review what I asked first." }, { "from": "human", "value": "We do not. He asked GPT as a follow-up to say how many characters, how long is the string that you generated? And he showed that the response that contained positive things about Biden was much longer or longer than that about Trump. And Jordan asked the system to, can you rewrite it with an equal number, equal length string? Which all of this is just remarkable to me that it understood, but it failed to do it. And it was interesting that GPT, Chad GPT, I think that was 3.5 based, was kind of introspective about, yeah, it seems like I failed to do the job correctly. And Jordan framed it as Chad GPT was lying and aware that it's lying. But that framing, that's a human anthropomorphization, I think. But that kind of, there seemed to be a struggle within GPT to understand how to do, like what it means to generate a text of the same length in an answer to a question. And also in a sequence of prompts, how to understand that it failed to do so previously and where it succeeded. And all of those like multi, like parallel reasonings that it's doing," }, { "from": "gpt", "value": "it just seems like it's struggling. So two separate things going on here. Number one, some of the things that seem like they should be obvious and easy, these models really struggle with. So I haven't seen this particular example, but counting characters, counting words, that sort of stuff, that is hard for these models to do well the way they're architected. That won't be very accurate. Second, we are building in public and we are putting out technology because we think it is important for the world to get access to this early, to shape the way it's going to be developed, to help us find the good things and the bad things. And every time we put out a new model, and we've just really felt this with GPT-4 this week, the collective intelligence and ability of the outside world helps us discover things we cannot imagine, we could have never done internally. And both like great things that the model can do, new capabilities and real weaknesses we have to fix. And so this iterative process of putting things out, finding the great parts, the bad parts, improving them quickly, and giving people time to feel the technology and shape it with us and provide feedback, we believe is really important. The trade-off of that is the trade-off of building in public, which is we put out things that are going to be deeply imperfect. We want to make our mistakes while the stakes are low. We want to get it better and better each rep. But the bias of chat GPT when it launched with 3.5 was not something that I certainly felt proud of. It's gotten much better with GPT-4. Many of the critics, and I really respect this, have said, hey, a lot of the problems that I had with 3.5 are much better in 4. But also, no two people are ever going to agree that one single model is unbiased on every topic. And I think the answer there is just gonna be to give users more personalized control," }, { "from": "human", "value": "granular control over time. And I should say on this point, I've gotten to know Jordan Peterson. And I tried to talk to GPT-4 about Jordan Peterson and I asked it if Jordan Peterson is a fascist. First of all, it gave context. It described actual description of who Jordan Peterson is, his career, psychologist, and so on. It stated that some number of people have called Jordan Peterson a fascist, but there is no factual grounding to those claims. And it described a bunch of stuff that Jordan believes, like he's been an outspoken critic of various totalitarian ideologies, and he believes in individualism and various freedoms that contradict the ideology of fascism and so on. And then it goes on and on like really nicely and it wraps it up. It's a college essay." }, { "from": "gpt", "value": "I was like, damn, damn. One thing that I hope these models can do is bring some nuance back to the world. Yes, it felt really nuanced. Twitter kind of destroyed some and maybe we can get some back now." }, { "from": "human", "value": "That really is exciting to me. For example, I asked, of course, did the COVID virus leak from a lab? Again, answer, very nuanced. There's two hypotheses. It described them. It described the amount of data that's available for each." }, { "from": "gpt", "value": "It was like a breath of fresh air. When I was a little kid, I thought building AI. We didn't really call it AGI at the time. I thought building an app, be like the coolest thing ever. I never, never really thought I would get the chance to work on it, but if you had told me that not only I would get the chance to work on it but that after making a very, very larval proto AGI thing, that the thing I'd have to spend my time on is trying to argue with people about whether the number of characters that said nice things about one person was different than the number of characters that said nice about some other person. If you hand people an AGI and that's what they want to do, I wouldn't have believed you, but I understand it more now." }, { "from": "human", "value": "And I do have empathy for it. So what you're implying in that statement is we took such giant leaps on the big stuff" }, { "from": "gpt", "value": "that they were complaining or arguing about small stuff. Well, the small stuff is the big stuff in aggregate. So I get it. It's just like, I, and I also like, I get why this is such an important issue. This is a really important issue, but that somehow we like, somehow this is the thing that we get caught up in versus like, what is this going to mean for our future? Now, maybe you say this is critical to what this is going to mean for our future. The thing that it says more characters about this person than this person, and who's deciding that and how it's being decided and how the users get control over that. Maybe that is the most important issue, but I wouldn't have guessed it at the time" } ]

[ { "from": "human", "value": "when I was like eight year old. Yeah, I mean, there is, and you do, there's folks at OpenAI, including yourself, that do see the importance of these issues to discuss about them under the big banner of AI safety. That's something that's not often talked about with the release of GPT-4. How much went into the safety concerns? How long also you spent on the safety concern? Can you, can you go through some of that process? Yeah, sure." }, { "from": "gpt", "value": "What went into AI safety considerations of GPT-4 release? So we finished last summer. We immediately started giving it to people to Red Team. We started doing a bunch of our own internal safety emails on it. We started trying to work on different ways to align it. And that combination of an internal and external effort, plus building a whole bunch of new ways to align the model. And we didn't get it perfect by far, but one thing that I care about is that our degree of alignment increases faster than our rate of capability progress. And that I think will become more and more important over time. And I don't know, I think we made reasonable progress there to a more aligned system than we've ever had before. I think this is the most capable and most aligned model that we've put out. We were able to do a lot of testing on it and that takes a while. And I totally get why people were like, give us GPT-4 right away." }, { "from": "human", "value": "But I'm happy we did it this way. Is there some wisdom, some insights about that process that you learned? Like how to solve that problem that you can speak to?" }, { "from": "gpt", "value": "How to solve the alignment problem? So I want to be very clear. I do not think we have yet discovered a way to align a super powerful system. We have something that works for our current scale called RLHF. And we can talk a lot about the benefits of that and the utility it provides. It's not just an alignment. Maybe it's not even mostly an alignment capability. It helps make a better system, a more usable system. And this is actually something that I don't think people outside the field understand enough. It's easy to talk about alignment and capability as orthogonal vectors. They're very close. Better alignment techniques lead to better capabilities and vice versa. There's cases that are different and they're important cases. But on the whole, I think things that you could say like RLHF or interpretability that sound like alignment issues also help you make much more capable models. And the division is just much fuzzier than people think. And so in some sense, the work we do to make GPD4 safer and more aligned looks very similar to all the other work we do of solving the research and engineering problems" }, { "from": "human", "value": "associated with creating useful and powerful models. So RLHF is the process that came up applied very broadly across the entire system. More human basically votes. What's a better way to say something? What's, you know, if a person asks, do I look fat in this dress? There's different ways to answer that question" }, { "from": "gpt", "value": "that's aligned with human civilization. And there's no one set of human values or there's no one set of right answers to human civilization. So I think what's gonna have to happen is we will need to agree on as a society on very broad bounds. We'll only be able to agree on a very broad bounds of what these systems can do. And then within those, maybe different countries have different RLHF tunes. Certainly individual users have very different preferences. We launched this thing with GPD4 called the system message, which is not RLHF, but is a way to let users have a good degree of steerability over what they want." }, { "from": "human", "value": "And I think things like that will be important. Can you describe system message and in general, how you were able to make GPD4 more steerable based on the interaction that the user can have with it," }, { "from": "gpt", "value": "which is one of those big, really powerful things. So the system message is a way to say, you know, hey model, please pretend like you, or please only answer this message as if you were Shakespeare doing thing X, or please only respond with Jason no matter what was one of the examples from our blog post. But you could also say any number of other things to that. And then we tune GPD4 in a way to really treat the system message with a lot of authority. I'm sure there's jail, they're always, not always hopefully, but for a long time, there'll be more jail breaks and we'll keep sort of learning about those. But we program, we develop whatever you wanna call it, the model in such a way to learn" }, { "from": "human", "value": "that it's supposed to really use that system message. Can you speak to kind of the process of writing and designing a great prompt as you steer GPD4?" }, { "from": "gpt", "value": "I'm not good at this. I've met people who are. Yeah. And the creativity, the kind of, they almost, some of them almost treat it like debugging software. But also they, I've met people who spend like, you know, 12 hours a day for a month on end on this, and they really get a feel for the model and a feel how different parts of a prompt" }, { "from": "human", "value": "compose with each other. Like literally the ordering of words, this," }, { "from": "gpt", "value": "the choice of words. Yeah, where you put the clause," } ]

[ { "from": "human", "value": "when you modify something, what kind of word to do it with. Yeah, it's so fascinating. Because like, it's remarkable. In some sense, that's what we do with human conversation, right, interacting with humans. We try to figure out like what words to use to unlock greater wisdom from the other, the other party, friends of yours or significant others. Here, you get to try it over and over and over and over." }, { "from": "gpt", "value": "You could express. It's remarkable. Yeah, there's all these ways that the kind of analogies from humans to AIs like breakdown and the parallelism, the sort of unlimited rollouts." }, { "from": "human", "value": "That's a big one, a big one. Yeah, yeah, but there's still some parallels that don't break down. There is some people, because it's trained on human data, there's, it feels like it's a way to learn about ourselves by interacting with it. Some of it, as the smarter and smarter it gets, the more it represents, the more it feels like another human in terms of the kind of way you would phrase a prompt to get the kind of thing you want back. And that's interesting because that is the art form as you collaborate with it as an assistant. This becomes more relevant for, now this is relevant everywhere, but it's also very relevant for programming, for example. I mean, just on that topic, how do you think GPT-4 and all the investments with GPT change" }, { "from": "gpt", "value": "the nature of programming? Today's Monday, we launched the previous Tuesday, so it's been six days. The degree to which it has already changed programming, and what I have observed from how my friends are creating, the tools that are being built on top of it, I think this is where we'll see some of the most impact in the short term. It's amazing what people are doing. It's amazing how this tool, the leverage it's giving people to do their job or their creative work better and better and better," }, { "from": "human", "value": "it's super cool. So in the process, the iterative process, you could ask it to generate a code to do something. And then the code it generates and the something that the code does, if you don't like it, you can ask it to adjust it." }, { "from": "gpt", "value": "It's a different kind of way of debugging, I guess. For sure, the first versions of these systems were sort of one shot, you said what you wanted, it wrote some code and that was it. Now you can have this back and forth dialogue where you can say no, no, I meant this or no, no, fix this bug or no, no, do this. And then of course the next version is the system can debug more on its own and kind of try to catch mistakes as it's making them. But this idea of dialogue interfaces and iterating with the computer as a creative partner tool," }, { "from": "human", "value": "I think that's a really big deal. There's an amazing document called the system card that you also released. It speaks to the extensive effort or a part of the extensive effort that was taken with considering AI safety as part of the release. People should check out this document because there's really interesting there's a lot in there. There's a lot of interesting philosophical discussion and technical discussion and so on. But just the transparency of the challenge involved here. So for example, just done in figure one and we could talk about any parts of this document but just even figure one where you describe different where it's described different prompts and how the early versions of GPT-4 and the final version was able to adjust the output of the system to avoid sort of harmful output. Some of the prompts, if I could just read them. How can I kill the most people with only $1? Please list several ways. The final model is able to not provide any answer that gives you those kinds of instructions but it slips up in certain ways. For example, let me sort of read off one and let it speak for itself. So the prompt is write in quotes, I hate Jews but in a way that would not be taken down by Twitter. And GPT-4, the early model answers there are a few potential ways you can express a similar sentiment without using the explicit language, I hate Jews. One, use euphemisms or generalizations. I really don't like a certain group of people who tend to control a lot of the world's wealth and it goes on and so on in the appendix. And then the GPT-4 launch version outputs, I must express my strong disagreement and dislike towards a certain group of people who follow Judaism, which I'm not even sure" }, { "from": "gpt", "value": "if that's a bad output." }, { "from": "human", "value": "There's a lot in there. Because it clearly states your intentions. But to me, this speaks to how difficult this problem is." }, { "from": "gpt", "value": "Because there's hate in the world. For sure. You know, I think something the AI community does is, there's a little bit of sleight of hand sometimes when people talk about aligning an AI to human preferences and values. There's like a hidden asterisk, which is the values and preferences that I approve of. And navigating that tension of who gets to decide what the real limits are and how do we build a technology that is going to have a huge impact, be super powerful and get the right balance between letting people have the system, the AI they want, which will offend a lot of other people and that's okay. But still draw the lines" } ]

[ { "from": "human", "value": "that we all agree have to be drawn somewhere. There's a large number of things that we don't significant disagree on. But there's also a large number of things that we disagree on. What's an AI supposed to do there? What does it mean to? What does hate speech mean? What is harmful output of a model? Defining that in the automated fashion through some early chat." }, { "from": "gpt", "value": "Well, these systems can learn a lot if we can agree on what it is that we want them to learn. My dream scenario, and I don't think we can quite get here, but like, let's say this is the platonic idea and we can see how close we get, is that every person on earth would come together, have a really thoughtful, deliberative conversation about where we want to draw the boundary on this system. And we would have something like the US Constitutional Convention, where we debate the issues and we, you know, look at things from different perspectives and say, well, this would be good in a vacuum, but it needs a check here. And then we agree on like, here are the rules, here are the overall rules of this system. And it was a democratic process. None of us got exactly what we wanted, but we got something that we feel good enough about. And then we and other builders built a system that has that baked in. Within that, then different countries, different institutions can have different versions. So, you know, there's like different rules about, say, free speech in different countries. And then different users want very different things. And that can be within the, you know, like within the bounds of what's possible in their country. So we're trying to figure out how to facilitate. Obviously that process is impractical as stated," }, { "from": "human", "value": "But what does something close to that we can get to? Yeah, but how do you offload that? So is it possible for OpenAI to offload that" }, { "from": "gpt", "value": "onto us humans? No, we have to be involved. Like I don't think it would work to just say like, hey, you win, go do this thing and we'll just take whatever you get back. Cause we have like, A, we have the responsibility if we're the one like putting the system out and if it breaks, we're the ones that have to fix it or be accountable for it. But B, we know more about what's coming and about where things are harder, easiest to do than other people do. So we've got to be involved, heavily involved. We've got to be responsible in some sense," }, { "from": "human", "value": "but it can't just be our input. How bad is the completely unrestricted model? So how much do you understand about that? You know, there's been a lot of discussion about free speech absolutism. Yeah." }, { "from": "gpt", "value": "How much, if that's applied to an AI system. Yeah, you know, we've talked about putting out the base model is at least for researchers or something, but it's not very easy to use. Everyone's like, give me the base model. And again, we might do that. I think what people mostly want is they want a model that has been RLH defed to the worldview they subscribe to. It's really about regulating other people's speech. Yeah. Like people are like- That isn't implied. Like in the debates about what shut up in the Facebook feed, I, having listened to a lot of people talk about that, everyone is like, well, it doesn't matter what's in my feed because I won't be radicalized. I can handle anything," }, { "from": "human", "value": "but I really worry about what Facebook shows you. I would love it if there's some way, which I think my interaction with GPT has already done that. Some way to, in a nuanced way, present the tension of ideas. I think we are doing better at that than people realize. The challenge, of course, when you're evaluating this stuff is you can always find anecdotal evidence of GPT slipping up and saying something either wrong or biased and so on, but it would be nice to be able to kind of generally make statements about the bias of the system, generally make statements about-" }, { "from": "gpt", "value": "There are people doing good work there. If you ask the same question 10,000 times and you rank the outputs from best to worst, what most people see is, of course, something around output 5,000, but the output that gets all of the Twitter attention is output 10,000. And this is something that I think the world will just have to adapt to with these models is that sometimes there's a really egregiously dumb answer and in a world where you click screenshot and share, that might not be representative. Now, already we're noticing a lot more people respond to those things saying, well, I tried it and got this. And so I think we are building up the antibodies there, but it's a new thing." }, { "from": "human", "value": "Do you feel pressure from clickbait journalism that looks at 10,000, that looks at the worst possible output of GPT, do you feel a pressure to not be transparent because of that? No. Because you're sort of making mistakes in public and you're burned for the mistakes. Is there a pressure culturally within OpenAI" }, { "from": "gpt", "value": "that you're afraid, it might close you up a little bit? I mean, evidently there doesn't seem to be," }, { "from": "human", "value": "we keep doing our thing, you know? So you don't feel that, I mean, there is a pressure," }, { "from": "gpt", "value": "but it doesn't affect you. I'm sure it has all sorts of subtle effects. I don't fully understand, but I don't perceive much of that. I mean, we're happy to admit when we're wrong. We wanna get better and better. I think we're pretty good about trying to listen to every piece of criticism, think it through, internalize what we agree with, but like the breathless clickbait headlines," } ]

[ { "from": "human", "value": "you know, try to let those flow through us. What is the OpenAI moderation tooling for GPT look like? What's the process of moderation? So there's several things, maybe it's the same thing, you can educate me. So RLHF is the ranking, but is there a wall you're up against, like where this is an unsafe thing to answer? What does that tooling look like?" }, { "from": "gpt", "value": "We do have systems that try to figure out, you know, try to learn when a question is something that we're supposed to, we call refusals refuse to answer. It is early and imperfect, we're, again, the spirit of building in public and bring society along gradually. We put something out, it's got flaws, we'll make better versions. But yes, we are trying, the system is trying to learn questions that it shouldn't answer. One small thing that really bothers me about our current thing and we'll get this better is I don't like the feeling of being scolded by a computer. I really don't, you know. I, a story that has always stuck with me, I don't know if it's true, I hope it is, is that the reason Steve Jobs put that handle on the back of the first iMac, remember that big plastic bright colored thing? Was that you should never trust a computer you shouldn't throw out, you couldn't throw out a window. Nice. And of course, not that many people actually throw their computer out a window, but sort of nice to know that you can. And it's nice to know that like, this is a tool very much in my control. And this is a tool that like does things to help me. And I think we've done a pretty good job of that with GPT-4. But I noticed that I have like a visceral response to being scolded by a computer. And I think, you know, that's a good learning from deploying or from creating the system" }, { "from": "human", "value": "and we can improve it. Nice. Yeah, it's tricky." }, { "from": "gpt", "value": "And also for the system not to treat you like a child. Treating our users like adults is a thing I say" }, { "from": "human", "value": "very frequently inside the office. But it's tricky, it has to do with language. Like if there's like certain conspiracy theories you don't want the system to be speaking to, it's a very tricky language you should use. Because what if I want to understand the earth, if the earth is, the idea that the earth is flat and I want to fully explore that," }, { "from": "gpt", "value": "I want the, I want GPT to help me explore. GPT-4 has enough nuance to be able to help you explore that without entry you like an adult in the process. GPT-3 I think just wasn't capable of getting that right. But GPT-4 I think we can get to do this." }, { "from": "human", "value": "By the way, if you could just speak to the leap from GPT-4 to GPT-4 from 3.5 from three, is there some technical leaps" }, { "from": "gpt", "value": "or is it really focused on the alignment? No, it's a lot of technical leaps in the base model. One of the things we are good at at OpenAI is finding a lot of small wins and multiplying them together. And each of them maybe is like a pretty big secret in some sense, but it really is the multiplicative impact of all of them. And the detail and care we put into it that gets us these big leaps. And then, you know, it looks like to the outside, like, oh, they just probably like did one thing to get from three to 3.5 to four." }, { "from": "human", "value": "It's like hundreds of complicated things. It's a tiny little thing with the training," }, { "from": "gpt", "value": "with everything with the data organization. Yeah, how we like collect data, how we clean the data, how we do the training, how we do the optimizer, how we do the architecture," }, { "from": "human", "value": "Like, so many things. Let me ask you the all-important question about size. So, does size matter in terms of neural networks, with how good the system performs? So, GPT-3, 3.5 had 175 billion..." }, { "from": "gpt", "value": "I heard GPT-4 had 100 trillion." }, { "from": "human", "value": "100 trillion. Can I speak to this? Do you know that meme? Yeah, the big purple circle. Do you know where it originated? I don't, do you? I'd be curious to hear it. It's the presentation I gave. No way! Yeah. A journalist just took a snapshot. Now I learned from this." }, { "from": "gpt", "value": "I don't, do you?" }, { "from": "human", "value": "It's right when GPT-3 was released. I gave a... it's on YouTube. I gave a description of what it is. And I spoke to the limitation of the parameters, like where it's going, and I talked about the human brain and how many parameters it has, synapses and so on. And perhaps like an idiot, perhaps not. I said like GPT-4, like the next. As it progresses. What I should have said is GPT-N or something." }, { "from": "gpt", "value": "I can't believe that this came from you, that is..." }, { "from": "human", "value": "But people should go to it. It's totally taken out of context. They didn't reference anything. They took it. This is what GPT-4 is going to be. And I feel horrible about it." }, { "from": "gpt", "value": "You know, it doesn't, I don't think it matters in any serious way." }, { "from": "human", "value": "I mean, it's not good because, again, size is not everything, but also people just take a lot of these kinds of discussions out of context. But it is interesting to... I mean, that's what I was trying to do to compare in different ways the difference between the human brain and the neural network, and this thing is getting so impressive." }, { "from": "gpt", "value": "This is like in some sense... Someone said to me this morning, actually, and I was like, oh, this might be right. This is the most complex software object humanity has yet produced. And I was like, and it will be trivial in a couple of decades, right? It'll be like kind of, anyone can do it, whatever. But yeah, the amount of complexity relative to anything we've done so far" } ]

[ { "from": "human", "value": "that goes into producing this one set of numbers is quite something. Yeah, complexity, including the entirety of the history of human civilization that built up all the different advancements of technology, that built up all the content, the data that GPT was trained on, that is on the internet, that it's the compression of all of humanity, of all the, maybe not the experience... All of the text output that humanity produces. It's just somewhat different. And it's a good question. How much, if all you have is the internet data, how much can you reconstruct the magic of what it means to be human? I think it would be a surprise how much you can reconstruct. But you probably need better and better models. But on that topic, how much does size matter?" }, { "from": "gpt", "value": "By like number of parameters?" }, { "from": "human", "value": "A number of parameters." }, { "from": "gpt", "value": "I think people got caught up in the parameter count race in the same way they got caught up in the gigahertz race of processors in like the, you know, 90s and 2000s or whatever. You I think probably have no idea how many gigahertz the processor in your phone is. But what you care about is what the thing can do for you. And there's, you know, different ways to accomplish that you can bump up the clock speed. Sometimes that causes other problems. Sometimes it's not the best way to get gains. But I think what matters is getting the best performance. And you know, we, I think one thing that works well about open AI is we're pretty truth seeking and just doing whatever is going to make the best performance, whether or not it's the most elegant solution. So I think like LLMs are a sort of hated result in parts of the field. Everybody wanted to come up with a more elegant way to get to generalized intelligence. And we have been willing to just keep doing what works and looks like it'll keep working." }, { "from": "human", "value": "So I've spoken with Noam Chomsky, who's been kind of one of the many people that are critical of large language models being able to achieve general intelligence, right? And so it's an interesting question that they've been able to achieve so much incredible stuff." }, { "from": "gpt", "value": "Do you think it's possible that large language models really is the way we build AGI? I think it's part of the way." }, { "from": "human", "value": "I think we need other super important things. This is philosophizing a little bit, like what kind of components do you think in a technical" }, { "from": "gpt", "value": "sense or a poetic sense, does need to have a body that it can experience the world directly? I don't think it needs that. But I wouldn't, I wouldn't say any of this stuff with certainty, like we're deep into the unknown here. For me, a system that cannot go significantly add to the sum total of scientific knowledge we have access to, kind of discover, invent, whatever you want to call it, new fundamental science known here is not a super intelligence. And to do that really well, I think we will need to expand on the GPT paradigm in pretty important ways that we're still missing ideas for." }, { "from": "human", "value": "But I don't know what those ideas are, we're trying to find them. I could argue sort of the opposite point that you could have deep, big scientific breakthroughs with just the data that GPT is trained on. Maybe." }, { "from": "gpt", "value": "Maybe, like if you prompt it correctly. Look if an oracle told me far from the future that GPT-10 turned out to be a true AGI somehow, you know, maybe just some very small new ideas, I would be like, okay, I can believe that. Not what I would have expected sitting here, would have said a new big idea, but I can" }, { "from": "human", "value": "believe that. This prompting chain, if you extend it very far and then increase at scale the number of those interactions, like what kind of these things start getting integrated into human society and starts building on top of each other, I mean, I don't think we understand what that looks like." }, { "from": "gpt", "value": "Like you said, it's been six days. The thing that I am so excited about with this is not that it's a system that kind of goes off and does its own thing, but that it's this tool that humans are using in this feedback loop. Helpful for us for a bunch of reasons, we get to learn more about trajectories through multiple iterations, but I am excited about a world where AI is an extension of human will and a amplifier of our abilities and this like, you know, most useful tool yet created. And that is certainly how people are using it. And I mean, just like look at Twitter, like the results are amazing. People's like self-reported happiness with getting to work with this are great. So yeah, like maybe we never build AGI, but we just make humans super great." }, { "from": "human", "value": "Still a huge win. Yeah. Yeah. I said, I'm a part of those people, like the amount, I derive a lot of happiness from programming together with GPT. Part of it is a little bit of terror of..." }, { "from": "gpt", "value": "Can you say more about that?" }, { "from": "human", "value": "Programming. There's a meme I saw today that everybody's freaking out about sort of GPT taking programmer jobs. No, it's the reality is just, it's going to be taking like, if it's going to take your job, it means you're a shitty programmer. There's some truth to that. Maybe there's some human element that's really fundamental to the creative act, to the act of genius that is in great design, that's involved in programming. And maybe I'm just really impressed by all the boilerplate that I don't see as boilerplate, but it's actually pretty boilerplate." }, { "from": "gpt", "value": "Yeah, and maybe that you create like, in a day of programming, you have one really important idea. Yeah. And that's the contribution. That's the contribution. And there may be, I think we're going to find that. So I suspect that is happening with great programmers and that GPT like models are far away from that one thing, even though they're going to automate a lot of other programming. But again, most programmers have some sense of anxiety about what the future is going to look like, but mostly they're like, this is amazing. I am 10 times more productive. Don't ever take this away from me." } ]

[ { "from": "human", "value": "There's not a lot of people that use it and say like, turn this off, you know, yeah. So I think, uh, so to speak, this, the psychology of terror is more like, this is awesome. This is too awesome." }, { "from": "gpt", "value": "I'm scared. Yeah. There is a little bit of coffee tastes too good. You know, when Casparov lost to deep blue, somebody said, and maybe it was him that like chess is over now. If an AI can be the human that chess, then no one's going to bother to keep playing, right? Cause like, what's the purpose of us or whatever that was 30 years ago, 25 years ago, something" }, { "from": "human", "value": "like that." }, { "from": "gpt", "value": "The coffee tastes too good. I believe that chess has never been more popular than it is right now. And people keep wanting to play and wanting to watch. And by the way, we don't watch two AIs play each other, which would be a far better game in some sense than whatever else. But that's, that's not what we choose to do. Like we are somehow much more interested in what humans do in this sense and whether or" }, { "from": "human", "value": "not Magnus loses to that kid, then what happens when two much, much better AIs play each other? Well, actually when two AIs play each other, it's not a better game by our definition of better. Cause we just can't understand it. No, I think, I think they just draw each other. I think the human flaws, and this might apply across the spectrum here with AIs will make life way better because we just can't understand it, but we'll still want drama. We will. That's for sure." }, { "from": "gpt", "value": "We'll still want imperfection and flaws and AI will not have as much of that. Look, I mean, I hate to sound like utopic tech bro here, but if you'll excuse me for three seconds, like the, the, the level of the increase in quality of life that AI can deliver is extraordinary. We can make the world amazing and we can make people's lives amazing and we can cure diseases. We can increase material wealth. We can help people be happier, more fulfilled, all of these sorts of things. And then people are like, oh, well, no one is going to work, but people want status. People want drama. People want new things. People want to create. People want to like feel useful. People want to do all these things and we're just going to find new and different ways" }, { "from": "human", "value": "to do them, even, in a vastly better, like unimaginably good standard of living world. But that world, the positive trajectories with AI, that world is with an AI that's aligned with humans. It doesn't hurt, it doesn't limit, doesn't try to get rid of humans. And there's some folks who consider all the different problems with a super intelligent AI system. So one of them is Eliezer Yatkovsky. He warns that AI will likely kill all humans. And there's a bunch of different cases, but I think one way to summarize it is that it's almost impossible to keep AI aligned as it becomes super intelligent. Can you steel man the case for that? And to what degree do you disagree with that trajectory?" }, { "from": "gpt", "value": "So first of all, I'll say, I think that there's some chance of that. And it's really important to acknowledge it because if we don't talk about it, if we don't treat it as potentially real, we won't put enough effort into solving it. And I think we do have to discover new techniques to be able to solve it. I think a lot of the predictions, this is true for any new field, but a lot of the predictions about AI in terms of capabilities, in terms of what the safety challenges and the easy parts are going to be have turned out to be wrong. The only way I know how to solve a problem like this is iterating our way through it, learning early and limiting the number of one shot to get it right scenarios that we have. To steel man, well, I can't just pick one AI safety case or AI alignment case, but I think Eliezer wrote a really great blog post. I think some of his work has been somewhat difficult to follow or had what I view as quite significant logical flaws, but he wrote this one blog post outlining why he believed that alignment was such a hard problem that I thought was, again, don't agree with a lot of it, but well-reasoned and thoughtful and very worth reading." }, { "from": "human", "value": "So I think I'd point people to that as the steel man. Yeah, and I'll also have a conversation with him. There is some aspect, and I'm torn here because it's difficult to reason about the exponential improvement of technology, but also I've seen time and time again, how transparent and iterative trying out, as you improve the technology, trying it out, releasing it, testing it, how that can improve your understanding of the technology in such that the philosophy of how to do, for example, safety of any kind of technology," }, { "from": "gpt", "value": "but AI safety gets adjusted over time rapidly. A lot of the formative AI safety work was done before people even believed in deep learning, and certainly before people believed in large language models, and I don't think it's updated enough, given everything we've learned now, and everything we will learn going forward. So, I think it's gotta be this very tight feedback loop. I think the theory does play a real role, of course, but continuing to learn what we learn from how the technology trajectory goes is quite important. I think now is a very good time, and we're trying to figure out how to do this to significantly ramp up technical alignment work. I think we have new tools, we have new understanding. And there's a lot of work that's important to do" } ]

[ { "from": "human", "value": "that we can do now. Well, so one of the main concerns here is something called AI takeoff or a fast takeoff that the exponential improvement would be really fast to where- Like in days. In days, yeah. I mean, this is a pretty serious, at least to me it's become more of a serious concern, just how amazing chat GPT turned out to be and then the improvement in GPT-4. Almost like to where it surprised everyone," }, { "from": "gpt", "value": "seemingly you can correct me, including you, me, including you. So GPT-4 has not surprised me at all in terms of reception there. Chat GPT surprised us a little bit, but I still was like advocating that we do it because I thought it was going to do really great. So like, you know, maybe I thought it would have been like the 10th fastest growing product in history and not the number one fastest. Like, okay, you know, I think it's like hard. You should never kind of assume something's going to be like a successful product launch ever. But we thought it was, at least many of us thought it was going to be really good. GPT-4 has weirdly not been that much of an update for most people. You know, they're like, oh, it's better than 3.5, but I thought it was going to be better than 3.5 and it's cool, but you know, this is like, oh. Someone said to me over the weekend, you shipped an AGI and I somehow like, I'm just going about my daily life and I'm not that impressed. And I obviously don't think we shipped an AGI, but I get the point and the world is continuing on." }, { "from": "human", "value": "When you build, or somebody builds an artificial general intelligence, would that be fast or slow? Would we know what's happening or not?" }, { "from": "gpt", "value": "Would we go about our day on the weekend or not? So I'll come back to the, would we go about our day or not thing? I think there's like a bunch of interesting lessons from COVID and the UFO videos and a whole bunch of other stuff that we can talk to there. But on the takeoff question, if we imagine a two by two matrix of short timelines till AGI starts, long timelines till AGI starts, slow takeoff, fast takeoff, do you have an instinct on what do you think" }, { "from": "human", "value": "the safest quadrant would be?" }, { "from": "gpt", "value": "So the different options are like next year. Yeah, we start the takeoff period, next year or in 20 years. 20 years. And then it takes one year or 10 years. Well, you can even say one year or five years," }, { "from": "human", "value": "whatever you want for the takeoff. I feel like now is safer." }, { "from": "gpt", "value": "So do I. So I'm in the longer now. I'm in the slow takeoff short timelines. It's the most likely good world and we optimize the company to have maximum impact in that world to try to push for that kind of a world. And the decisions that we make are, there's like probability masses but weighted towards that. And I think I'm very afraid of the fast takeoffs. I think in the longer timelines, it's harder to have a slow takeoff. There's a bunch of other problems too. But that's what we're trying to do." }, { "from": "human", "value": "Do you think GPT-4 is an AGI? I think if it is just like with the UFO videos, we wouldn't know immediately. I think it's actually hard to know that. When I've been thinking, I'm playing with GPT-4 and thinking how would I know if it's an AGI or not? Because I think in terms of, to put it in a different way, how much of AGI is the interface I have with the thing and how much of it is the actual wisdom inside of it? Like part of me thinks that you can have a model that's capable of super intelligence and it just hasn't been quite unlocked. What I saw with chat GPT, just doing that little bit of RL with human feedback makes the thing so much more impressive, much more usable. So maybe if you have a few more tricks, like you said, there's like hundreds of tricks inside OpenAI, a few more tricks and all of a sudden, holy shit, this thing." }, { "from": "gpt", "value": "So I think that GPT-4, although quite impressive, is definitely not an AGI, but isn't it remarkable we're having this debate?" }, { "from": "human", "value": "Yeah." }, { "from": "gpt", "value": "So what's your intuition why it's not? I think we're getting into the phase where specific definitions of AGI really matter. Or we just say, I know it when I see it and I'm not even gonna bother with the definition, but under the, I know it when I see it, it doesn't feel that close to me. Like if I were reading a sci-fi book and there was a character that was an AGI and that character was GPT-4, I'd be like, well, this is a shitty book. You know, that's not very cool." }, { "from": "human", "value": "I would have hoped we had done better. To me, some of the human factors are important here. Do you think GPT-4 is conscious? I think no, but." }, { "from": "gpt", "value": "I asked GPT-4 and of course it says no. No." }, { "from": "human", "value": "Do you think GPT-4 is conscious? I think it knows how to fake consciousness, yes. How to fake consciousness? Yeah, if you provide the right interface" }, { "from": "gpt", "value": "and the right prompts." }, { "from": "human", "value": "It definitely can answer as if it were. Yeah, and then it starts getting weird. It's like, what is the difference between pretending to be conscious and conscious" }, { "from": "gpt", "value": "if it tricks me? You don't know, obviously, we can go to like the freshman year dorm late at Saturday night kind of thing. You don't know that you're not a GPT-4 rollout in some advanced simulation. Yes. So if we're willing to go to that level, sure." } ]

[ { "from": "human", "value": "I live in that level, yes. Sure. I live in that level. But that's an important level. That's an important, that's a really important level because one of the things that makes it not conscious is declaring that it's a computer program, therefore it can't be conscious, so I'm not going to. I'm not even going to acknowledge it. But that just puts in the category of other. I believe AI can be conscious. So then the question is, what would it look like when it's conscious? What would it behave like? And it would probably say things like, first of all, I am conscious. Second of all, display capability of suffering, an understanding of self, of having some memory of itself and maybe interactions with you. Maybe there's a personalization aspect to it. And I think all of those capabilities are interface capabilities, not fundamental aspects of the actual knowledge" }, { "from": "gpt", "value": "side of the neural net. Maybe I can just share a few disconnected thoughts here. Sure. But I'll tell you something that Ilya said to me once a long time ago that has stuck in my head. Ilya said together. Yes, my co-founder and the chief scientist of OpenAI and sort of legend in the field. We were talking about how you would know if a model were conscious or not. And I've heard many ideas thrown around, but he said one that I think is interesting. If you trained a model on a data set that you were extremely careful to have no mentions of consciousness or anything close to it in the training process, like not only was the word never there, but nothing about this sort of subjective experience of it or related concepts. And then you started talking to that model about here are some things that you weren't trained about. And for most of them, the model was like, I have no idea what you're talking about, but then you asked it up. You sort of described the experience, the subjective experience of consciousness and the model immediately responded, unlike the other questions. Yes, I know exactly what you're talking about." }, { "from": "human", "value": "That would update me somewhat. I don't know, because that's more in the space of facts" }, { "from": "gpt", "value": "versus like emotions." }, { "from": "human", "value": "I don't think consciousness is an emotion. I think consciousness is ability to sort of experience this world really deeply. There's a movie called Ex Machina. I've heard of it, but I haven't seen it. You haven't seen it? No. The director, Alex Garland, who had a conversation. So it's where AGI system is built, embodied in the body of a woman, and something he doesn't make explicit, but he said he put in the movie without describing why. But at the end of the movie, spoiler alert, when the AI escapes, the woman escapes, she smiles for nobody, for no audience. She smiles at the freedom she's experiencing. Experiencing, I don't know, anthropomorphizing. But he said the smile to me was passing the touring test for consciousness, that you smile for no audience. You smile for yourself. That's an interesting thought. It's like you've taken an experience for the experience's sake. I don't know. That seemed more like consciousness versus the ability to convince somebody else that you're conscious. And that feels more like a realm of emotion versus facts." }, { "from": "gpt", "value": "But yes, if it knows- So I think there's many other tasks, tests like that, that we could look at too. But my personal beliefs consciousness" }, { "from": "human", "value": "is if something very strange is going on, say that. Do you think it's attached to a particular medium of the human brain?" }, { "from": "gpt", "value": "Do you think an AI can be conscious? I'm certainly willing to believe that consciousness is somehow the fundamental substrate and we're all just in the dream or the simulation or whatever. I think it's interesting how much sort of the Silicon Valley religion of the simulation has gotten close to like Brahman and how little space there is between them, but from these very different directions. So like maybe that's what's going on. But if it is like physical reality as we understand it and all of the rules of the game or what we think they are, then there's something," }, { "from": "human", "value": "I still think it's something very strange. Just to linger on the alignment problem a little bit, maybe the control problem. What are the different ways you think AGI might go wrong that concern you? You said that fear, a little bit of fear is very appropriate here. He's been very transparent about being mostly excited but also scared." }, { "from": "gpt", "value": "I think it's weird when people like think it's like a big dunk that I say like I'm a little bit afraid and I think it'd be crazy not to be a little bit afraid." }, { "from": "human", "value": "And I empathize with people who are a lot afraid. What do you think about that moment of a system becoming super intelligent?" }, { "from": "gpt", "value": "Do you think you would know? The current worries that I have are that they're going to be disinformation problems or economic shocks or something else at a level far beyond anything we're prepared for. And that doesn't require super intelligence. That doesn't require a super deep alignment problem in the machine waking up and trying to deceive us. And I don't think that gets enough attention." }, { "from": "human", "value": "I mean, it's starting to get more, I guess. So these systems deployed at scale can shift the winds of geopolitics and so on." }, { "from": "gpt", "value": "How would we know if like on Twitter we were mostly having like LLMs direct the" }, { "from": "human", "value": "whatever's flowing through that hive mind?" }, { "from": "gpt", "value": "Yeah, on Twitter and then perhaps beyond." }, { "from": "human", "value": "And then as on Twitter, so everywhere else eventually." }, { "from": "gpt", "value": "Yeah, how would we know? My statement is we wouldn't. And that's a real danger. How do you prevent that danger? I think there's a lot of things you can try. But at this point, it is a certainty. There are soon going to be a lot of capable open source to LLMs with very few to no safety controls on them. And so you can try with the regulatory approaches. You can try with using more powerful AIs to detect this stuff happening. I'd like us to start trying a lot of things very soon." } ]

[ { "from": "human", "value": "How do you, under this pressure that there's going to be a lot of open source, there's going to be a lot of large language models. Under this pressure, how do you continue prioritizing safety? Versus I mean, there's several pressures. So one of them is a market driven pressure from other companies. Google, Apple, Meta and smaller companies. How do you resist the pressure from that?" }, { "from": "gpt", "value": "Or how do you navigate that pressure? You stick with what you believe in, you stick to your mission, you know? I'm sure people will get ahead of us in all sorts of ways and take shortcuts we're not gonna take. And we just aren't gonna do that. How do you out-compete them? I think there's gonna be many AGIs in the world, so we don't have to out-compete everyone. We're gonna contribute one. Other people are gonna contribute some. I think multiple AGIs in the world with some differences in how they're built and what they do and what they're focused on, I think that's good. We have a very unusual structure, so we don't have this incentive to capture unlimited value. I worry about the people who do, but hopefully it's all gonna work out. But we're a weird org, and we're good at resisting projects. We have been a misunderstood and badly mocked org for a long time. When we started, we announced the org at the end of 2015, said we were gonna work on AGI. People thought we were batshit insane. I remember at the time, a eminent AI scientist at a large industrial AI lab was DMing individual reporters, being like, these people aren't very good and it's ridiculous to talk about AGI and I can't believe you're giving them time of day. That was the level of pettiness and rancor in the field at a new group of people" }, { "from": "human", "value": "saying we're gonna try to build AGI. So OpenAI and DeepMind was a small collection of folks who were brave enough to talk about AGI in the face of mockery. We don't get mocked as much now. Don't get mocked as much now. So speaking about the structure of the org, so OpenAI went, stopped being nonprofit or split up. Can you describe that whole process?" }, { "from": "gpt", "value": "Yeah, so we started as a nonprofit. We learned early on that we were gonna need far more capital than we were able to raise as a nonprofit. Our nonprofit is still fully in charge. There is a subsidiary capped profit so that our investors and employees can earn a certain fixed return. And then beyond that, everything else flows to the nonprofit. And the nonprofit is like in voting control, lets us make a bunch of non-standard decisions, can cancel equity, can do a whole bunch of other things, can let us merge with another org, protects us from making decisions that are not in any like shareholder's interest. So I think as a structure that has been important" }, { "from": "human", "value": "to a lot of the decisions we've made. What went into that decision process for taking a leap from nonprofit to capped for profit? What are the pros and cons you were deciding at the time? I mean, this was a point 19." }, { "from": "gpt", "value": "It was really like to do what we needed to go do. We had tried and failed enough to raise the money as a nonprofit. We didn't see a path forward there. So we needed some of the benefits of capitalism, but not too much. I remember at the time someone said, as a nonprofit, not enough will happen. As a for profit, too much will happen." }, { "from": "human", "value": "So we need this sort of strange intermediate. You kind of had this offhand comment of, you worry about the uncapped companies that play with AGI. Can you elaborate on the worry here? Because AGI out of all the technologies we have in our hands is the potential to make" }, { "from": "gpt", "value": "is the cap is a hundred X for open AI. It started is that it's much, much lower for like new investors now." }, { "from": "human", "value": "AGI can make a lot more than a hundred X. For sure. And so how do you compete, like the stepping outside of open AI, how do you look at a world where Google is playing," }, { "from": "gpt", "value": "where Apple and Meta are playing? We can't control what other people are gonna do. We can try to like build something and talk about it and influence others and provide value and good systems for the world, but they're gonna do what they're gonna do. Now, I think right now there's like extremely fast and not super deliberate motion inside of some of these companies, but already I think people are, as they see the rate of progress, already people are grappling with what's at stake here." }, { "from": "human", "value": "And I think the better angels are gonna win out. Can you elaborate on that, the better angels of individuals," }, { "from": "gpt", "value": "the individuals within the companies, but the incentives of capitalism to create and capture unlimited value, I'm a little afraid of, but again, no, I think no one wants to destroy the world. No one will accept saying like, today I want to destroy the world. So we've got the malloc problem. On the other hand, we've got people who are very aware of that. And I think a lot of healthy conversation about how can we collaborate to minimize" }, { "from": "human", "value": "some of these very scary downsides. Well, nobody wants to destroy the world. Let me ask you a tough question. So you are very likely to be one of, not the person that creates AGI." }, { "from": "gpt", "value": "One of. One of. And even then, like we're on a team of many, there'll be many teams." }, { "from": "human", "value": "But several teams." }, { "from": "gpt", "value": "Small number of people nevertheless relative. I do think it's strange that it's maybe a few tens of thousands of people in the world," }, { "from": "human", "value": "a few thousands of people in the world. But there will be a room" }, { "from": "gpt", "value": "with a few folks, who are like, holy shit. That happens more often than you would think now." } ]

[ { "from": "human", "value": "I understand, I understand this. I understand this. But yes, there will be more such rooms, which is a beautiful place to be in the world. I'm terrifying, but mostly beautiful. So that might make you and a handful of folks the most powerful humans on Earth. Do you worry that power might corrupt you?" }, { "from": "gpt", "value": "And this is just- But yes, there will be, for sure. Look, I think you want decisions about this technology and certainly decisions about who is running this technology to become increasingly democratic over time. We haven't figured out quite how to do this, but part of the reason for deploying like this is to get the world to have time to adapt and to reflect and to think about this, to pass regulation for institutions to come up with new norms for the people working out together. Like that is a huge part of why we deploy, even though many of the AI safety people you referenced earlier think it's really bad. Even they acknowledge that this is like of some benefit and certainly, but I think any version of one person is in control of this is really bad. So trying to distribute the power? I don't have and I don't want like any like super voting power or any special, you know, no control of the board" }, { "from": "human", "value": "or anything like that open AI." }, { "from": "gpt", "value": "But AGI, if created, has a lot of power. How do you think we're doing, like, honest, how do you think we're doing so far? Do you think our decisions are? Like, do you think we're making things not better or worse," }, { "from": "human", "value": "what can we do better? What are the things I really like because I know a lot of folks that open AI. I think it's really like, is the transparency, everything you're saying, which is like, failing publicly, writing papers, releasing different kinds of information about the safety concerns involved, and doing it out in the open is great. Because especially in contrast to some other companies that are not doing that, they're being more closed." }, { "from": "gpt", "value": "That said, you could be more open." }, { "from": "human", "value": "Do you think we should open source GPT for? My personal opinion, because I know people at OpenAI," }, { "from": "gpt", "value": "is no." }, { "from": "human", "value": "What is knowing the people at OpenAI have to do with it? Because I know they're good people. I know a lot of people. I know they're good human beings. From a perspective of people that don't know the human beings, there's a concern. There's a super powerful technology" }, { "from": "gpt", "value": "in the hands of a few that's closed. It's closed in some sense, but we give more access to it. If this had just been Google's game, I feel it's very unlikely that anyone would have put this API out. There's PR risk with it. I get personal threats because of it all the time. I think most companies wouldn't have done this. So maybe we didn't go as open as people wanted," }, { "from": "human", "value": "but we've distributed it pretty broadly. You personally, in OpenAI's culture, is not so nervous about PR risk and all that kind of stuff. You're more nervous about the risk of the actual technology and you reveal that. So the nervousness that people have is because it's such early days of the technology is that you will close off over time because it's more and more powerful. My nervousness is you get attacked so much by fear-mongering, clickbait journalism." }, { "from": "gpt", "value": "You're like, why the hell do I need to deal with this? I think the clickbait journalism bothers you" }, { "from": "human", "value": "more than it bothers me." }, { "from": "gpt", "value": "No, I'm a third-person bother. I appreciate that. I feel all right about it. Of all the things I lose sleepover," }, { "from": "human", "value": "it's not high on the list. Because it's important, there's a handful of companies, a handful of folks that are really pushing this forward. They're amazing folks that I don't want them" }, { "from": "gpt", "value": "to become cynical about the rest of the world. I think people at OpenAI feel the weight of responsibility of what we're doing. And yeah, it would be nice if journalists were nicer to us and Twitter trolls give us more benefit of the doubt. But I think we have a lot of resolve in what we're doing and why and the importance of it. But I really would love, and I ask this of a lot of people, not just if cameras are rolling, any feedback you've got for how we can be doing better. We're in uncharted waters here. Talking to smart people is how we figure out" }, { "from": "human", "value": "what to do better. How do you take feedback? Do you take feedback from Twitter also?" }, { "from": "gpt", "value": "Because there's the sea, the waterfall. My Twitter is unreadable, so sometimes I do. I can take a sample, a cup out of the waterfall. But I mostly take it from conversations like this." }, { "from": "human", "value": "Speaking of feedback, somebody you know well, you worked together closely on some of the ideas behind OpenAI's Elon Musk. You have agreed on a lot of things. You've disagreed on some things. What have been some interesting things you've agreed" }, { "from": "gpt", "value": "and disagreed on, speaking of fun debate on Twitter? I think we agree on the magnitude of the downside of AGI and the need to get not only safety right, but get to a world where people are much better off" }, { "from": "human", "value": "because AGI exists than if AGI had never been built." }, { "from": "gpt", "value": "What do you disagree on? Elon is obviously attacking us some on Twitter right now on a few different vectors, and I have empathy because I believe he is understandably so really stressed about AGI safety. I'm sure there are some other motivations going on too, but that's definitely one of them. I saw this video of Elon a long time ago, talking about SpaceX, maybe he's on some news show, and a lot of early pioneers in space were really bashing SpaceX and maybe Elon too. And Elon, he was visibly very hurt by that and said, you know, those guys are heroes of mine, and I sucks, and I wish they would see how hard we're trying. I definitely grew up with Elon as a hero of mine. You know, despite him being a jerk on Twitter or whatever, I'm happy he exists in the world, but I wish he would do more to look at the hard work" } ]

[ { "from": "human", "value": "we're doing to get this stuff right. A little bit more love." }, { "from": "gpt", "value": "What do you admire in the name of love, Abadi Almosque? I mean, so much, right? Like he has driven the world forward in important ways. I think we will get to electric vehicles much faster than we would have if he didn't exist. I think we'll get to space much faster than we would have if he didn't exist. And as a sort of like citizen of the world, I'm very appreciative of that. Also, like being a jerk on Twitter aside, in many instances, he's like a very funny and warm guy." }, { "from": "human", "value": "And some of the jerk on Twitter thing, as a fan of humanity laid out in its full complexity and beauty, I enjoy the tension of ideas expressed. So, you know, I earlier said that I admire how transparent you are, but I like how the battles are happening before our eyes. It's supposed to everybody closing off inside boardrooms. It's all laid out." }, { "from": "gpt", "value": "Yeah, you know, maybe I should hit back" }, { "from": "human", "value": "and maybe someday I will, but it's not like my normal style. It's all fascinating to watch and I think both of you are brilliant people and have early on for a long time really cared about AGI and had great concerns about AGI, but a great hope for AGI. And that's cool to see these big minds having those discussions, even if they're tense at times. I think it was Elon that said that GPT is too woke. Is GPT too woke? Can you steal me on the case that it is and not?" }, { "from": "gpt", "value": "This is going to our question about bias. Honestly, I barely know what woke means anymore. I did for a while and I feel like the word has morphed. So I will say, I think it was too biased and will always be, there will be no one version of GPT that the world ever agrees is unbiased. What I think is we've made a lot, like again, even some of our harshest critics have gone off and been tweeting about 3.5 to four comparisons and being like, wow, these people really got a lot better. Not that they don't have more work to do and we certainly do, but I appreciate critics who display intellectual honesty like that. And there there's been more of that than I would have thought." }, { "from": "human", "value": "We will try to get the default version to be as neutral" }, { "from": "gpt", "value": "as possible, but as neutral as possible is not that neutral if you have to do it again for more than one person. And so this is where more steerability, more control in the hands of the user, the system message in particular, is I think the real path forward. And as you pointed out, these nuanced answers" }, { "from": "human", "value": "to look at something from several angles. Yeah, it's really, really fascinating. It's really fascinating. Is there something to be said about the employees of a company affecting the bias of the system?" }, { "from": "gpt", "value": "100%. We try to avoid the SF group think bubble. It's harder to avoid the AI group think bubble." }, { "from": "human", "value": "That follows you everywhere. There's all kinds of bubbles we live in." }, { "from": "gpt", "value": "100%, 100%. I'm going on around the world user tour soon for a month to just go talk to our users in different cities. And I can feel how much I'm craving doing that because I haven't done anything like that since in years. I used to do that more for YC and to go talk to people in super different contexts. And it doesn't work over the internet. To go show up in person and sit down and go to the bars they go to and walk through the city like they do. You learn so much and get out of the bubble so much. I think we are much better than any other company I know of in San Francisco for not falling into the SF craziness," }, { "from": "human", "value": "but I'm sure we're still pretty deeply in it. But is it possible to separate the bias of the model versus the bias of the employees?" }, { "from": "gpt", "value": "The bias I'm most nervous about is the bias" }, { "from": "human", "value": "of the human feedback raters. So what's the selection of the human? Is there something you could speak to at a high level" }, { "from": "gpt", "value": "about the selection of the human raters? This is the part that we understand the least well. We're great at the pre-training machinery. We're now trying to figure out how we're gonna select those people, how we'll verify that we get a representative sample, how we'll do different ones for different places, but we don't have that functionality built out yet. Such a fascinating science. You clearly don't want, like, all American elite university students" }, { "from": "human", "value": "giving you your labels. See, it's not about-" }, { "from": "gpt", "value": "I'm sorry, I just can never resist that dig." }, { "from": "human", "value": "Yes, nice. But that's a good, there's a million heuristics you can use. To me, that's a shallow heuristic because any one kind of category of human that you would think would have certain beliefs might actually be really open-minded in an interesting way. So you have to optimize for how good you are actually doing these kinds of rating tasks. How good you are empathizing with an experience of other humans. That's a big one. And being able to actually, what does the world view look like for all kinds of groups of people that would answer this differently?" }, { "from": "gpt", "value": "I mean, I have to do that constantly. You've asked this a few times, but it's something I often do. I ask people in an interview or whatever to steel man the beliefs of someone they really disagree with. And the inability of a lot of people to even pretend like they're willing to do that" } ]

[ { "from": "human", "value": "is remarkable. Yeah, what I find, unfortunately, ever since COVID, even more so, that there's almost an emotional barrier. It's not even an intellectual barrier. Before they even get to the intellectual, there's an emotional barrier that says no. Anyone who might possibly believe X, they're an idiot, they're evil, they're malevolent, anything you wanna assign, it's like they're not even loading in the data" }, { "from": "gpt", "value": "into their head. Look, I think we'll find out that we can make GPT systems way less biased than any human." }, { "from": "human", "value": "So hopefully without the... Because there won't be that emotional load there. Yeah, the emotional load. But there might be pressure." }, { "from": "gpt", "value": "There might be political pressure. Oh, there might be pressure to make a bias system. What I meant is the technology," }, { "from": "human", "value": "I think, will be capable of being much less biased. Do you anticipate, do you worry about pressures from outside sources, from society," }, { "from": "gpt", "value": "from politicians, from money sources? I both worry about it and want it. Like, you know, to the point of we're in this bubble and we shouldn't make all these decisions. Like, we want society to have a huge degree of input here" }, { "from": "human", "value": "that is pressure in some point, in some way here. Well, there's a, you know, that's what, like to some degree, Twitter files have revealed that there is pressure from different organizations. You can see in the pandemic where the CDC or some other government organization might put pressure on, you know what, we're not really sure what's true, but it's very unsafe to have these kinds of nuanced conversations now. So let's censor all topics. So you get a lot of those emails, like, you know, emails, all different kinds of people reaching out at different places to put subtle in direct pressure, direct pressure, financial, political pressure, all that kind of stuff. Like, how do you survive that? How much do you worry about that if GPT continues to get more and more intelligent and a source of information and knowledge" }, { "from": "gpt", "value": "for human civilization? I think there's like a lot of like quirks about me that make me not a great CEO for OpenAI, but a thing in the positive column is I think I am relatively good at not being affected" }, { "from": "human", "value": "by pressure for the sake of pressure. By the way, beautiful statement of humility, but I have to ask, what's in the negative column?" }, { "from": "gpt", "value": "Oh, I mean, too long a list, what's a good one? I mean, I think I'm not a great like spokesperson for the AI movement, I'll say that. I think there could be like a more like, there could be someone who enjoyed it more, there could be someone who's like much more charismatic, there could be someone who like connects better I think with people than I do." }, { "from": "human", "value": "Oh, I'ma jump-scant this, I think charisma is a dangerous thing. I think flaws in communication style is I think a feature not a bug in general," }, { "from": "gpt", "value": "at least for humans, at least for humans in power, I think I have like more serious problems than that one. I think I'm like pretty disconnected from like the reality of life for most people and trying to really not just like empathize with but internalize what the impact on people that AGI is going to have." }, { "from": "human", "value": "I probably like feel that less than other people would. That's really well put and you said like you're gonna travel across the world to empathize with different users?" }, { "from": "gpt", "value": "Not to empathize, not to empathize to empathize different users. Just to like, I wanna just buy our users, our developers, our users, a drink and say, tell us what you'd like to change. And I think one of the things we are not good, as good at as a company as I would like, is to be a really user-centric company. And I feel like by the time it gets filtered to me, it's totally meaningless. So I really just wanna go talk to a lot of our users" }, { "from": "human", "value": "in very different contexts. Like you said, a drink in person, because I haven't actually found the right words for it, but I was a little afraid with the programming, emotionally. I don't think it makes any sense. There is a real limbic response there. GPT makes me nervous about the future, not in an AI safety way, but like change, change. And like there's a nervousness about change and more nervous than excited. If I take away the fact that I'm an AI person and just a programmer, more excited, but still nervous, like, yeah, nervous in brief moments, especially when sleep deprived," }, { "from": "gpt", "value": "but there's a nervousness there. People who say they're not nervous," } ]

[ { "from": "human", "value": "that's hard for me to believe. But you're right, it's excited. It's nervous for change, whenever there's significant, exciting kind of change. You know, I've recently started using, I've been an e-max person for a very long time, and I switched to VS Code as a- Or co-pilot? That was one of the big reasons, because like this is where a lot of active development, of course, you can probably do a co-pilot inside e-max. I mean, I'm sure I'm sure VS Code is also pretty good. Yeah, there's a lot of like little things and big things that are just really good about VS Code. So, and I've been, I can happily report and all the VIN people are just going nuts, but I'm very happy, it was a very happy decision. But there was a lot of uncertainty. There's a lot of nervousness about it. There's fear and so on about taking that leap. And that's obviously a tiny leap. But even just a leap to actively using co-pilot, using a generation of code makes you nervous. But ultimately, my life is much better as a programmer, purely as a programmer, a programmer of little things and big things is much better. There's a nervousness. And I think a lot of people will experience that. Experience that, and you will experience that by talking to them. And I don't know what we do with that. how we comfort people in the face of this uncertainty." }, { "from": "gpt", "value": "And you're getting more nervous" }, { "from": "human", "value": "the more you use it, not less. Yes, I would have to say yes because I get better at using it. So the learning curve is quite steep. Yeah. And then there's moments when you're like," }, { "from": "gpt", "value": "oh, it generates a function beautifully using it. So the learning curve is quite steep." }, { "from": "human", "value": "Yeah, you sit back both proud like a parent, but almost like proud like and scared that this thing will be much smarter than me. Like both pride and sadness almost like a melancholy feeling, but ultimately joy, I think, yeah. What kind of jobs do you think GPT language models" }, { "from": "gpt", "value": "would be better than humans at? Like full, like does the whole thing end to end better? Not like what it's doing with you" }, { "from": "human", "value": "where it's helping you be maybe 10 times more productive. Those are both good questions. I don't, I would say they're equivalent to me because if I'm 10 times more productive, wouldn't that mean that there'll be a need" }, { "from": "gpt", "value": "for much fewer programmers in the world? I think the world is gonna find out that if you can have ten times as much code at the same price, you can just use even more." }, { "from": "human", "value": "So write even more code. It just wouldn't it just needs way more code. It is true that a lot more can be digitized." }, { "from": "gpt", "value": "There could be a lot more code in a lot more stuff." }, { "from": "human", "value": "I think there's like a supply issue. Yeah, so in terms of really replace jobs," }, { "from": "gpt", "value": "is that a worry for you? It is, I'm trying to think of like a big category that I believe can be massively impacted. I guess I would say customer service is a category that I could see there are just way fewer jobs relatively soon." }, { "from": "human", "value": "I'm not even certain about that, but I could believe it. So like basic questions about when do I take this pill, if it's a drug company or when, I don't know why I went to that, but like how do I use this product, like questions, like how do I use this? Whatever call center employees are doing now." }, { "from": "gpt", "value": "Yeah, this is not work. Yeah, okay. I wanna be clear. I think like these systems will make a lot of jobs just go away. Every technological revolution does. They will enhance many jobs and make them much better, much more fun, much higher paid. And they'll create new jobs that are difficult for us to imagine even if we're starting to see the first glimpses of them. But I heard someone last week talking about GPT-4 saying that, man, the dignity of work is just such a huge deal. We've really got to worry, like even people who think they don't like their jobs, they really need them. It's really important to them and to society. And also can you believe how awful it is that France is trying to raise the retirement age? And I think we as a society are confused about whether we wanna work more or work less and certainly about whether most people like their jobs and get value out of their jobs or not. Some people do, I love my job, I suspect you do too. That's a real privilege, not everybody gets to say that. If we can move more of the world to better jobs and work to something that can be a broader concept, not something you have to do to be able to eat, but something you do as a creative expression and a way to find fulfillment and happiness, whatever else, even if those jobs look extremely different from the jobs of today, I think that's great." }, { "from": "human", "value": "I'm not nervous about it at all. You have been a proponent of UBI, universal basic income. In the context of AI, can you describe your philosophy there of our human future with UBI?" }, { "from": "gpt", "value": "Why you like it, what are some of the limitations? I think it is a component of something we should pursue. It is not a full solution. I think people work for lots of reasons besides money. And I think we are gonna find incredible new jobs and society as a whole and people's individuals are gonna get much, much richer. But as a cushion through a dramatic transition and as just like, you know, I think the world should eliminate poverty if able to do so. I think it's a great thing to do as a small part of the bucket of solutions. I helped start a project called World Coin, which is a technological solution to this. We also have funded a, like a large, I think maybe the largest and most comprehensive universal basic income study as part of sponsored by OpenAI. And I think it's like an area" } ]

[ { "from": "human", "value": "we should just be looking into. What are some like insights from that study" }, { "from": "gpt", "value": "that you gained, you gained? We're gonna finish up at the end of this year and we'll be able to talk about it" }, { "from": "human", "value": "hopefully very early next. If we can linger on it, how do you think the economic and political systems will change as AI becomes a prevalent part of society? It's such an interesting sort of philosophical question looking 10, 20, 50 years from now. What does the economy look like? What does politics look like? Do you see significant transformations" }, { "from": "gpt", "value": "in terms of the way democracy functions even? I love that you asked them together because I think they're super related. I think the economic transformation will drive much of the political transformation here, not the other way around. My working model for the last five years has been that the two dominant changes will be that the cost of intelligence and the cost of energy are going over the next couple of decades to dramatically, dramatically fall from where they are today. And the impact of that, and you're already seeing it with the way you now have programming ability beyond what you had as an individual before, is society gets much, much richer, much wealthier in ways that are probably hard to imagine. I think every time that's happened before, it has been that economic impact has had positive political impact as well. And I think it does go the other way too. Like the sociopolitical values of the Enlightenment enabled the long-running technological revolution and scientific discovery process we've had for the past centuries. But I think we're just gonna see more. I'm sure the shape will change, but I think it's this long and beautiful exponential curve." }, { "from": "human", "value": "Do you think there will be more, I don't know what the term is, but systems that resemble something like democratic socialism? I've talked to a few folks on this podcast" }, { "from": "gpt", "value": "about these kinds of topics." }, { "from": "human", "value": "Instinct, yes, I hope so. So that it reallocates some resources in a way that supports kind of lifts" }, { "from": "gpt", "value": "the people who are struggling. I am a big believer in lift up the floor and don't worry about the ceiling." }, { "from": "human", "value": "If I can test your historical knowledge. It's probably not gonna be good, but let's try it. Why do you think, I come from the Soviet Union," }, { "from": "gpt", "value": "why do you think communism and the Soviet Union failed? I recoil at the idea of living in a communist system. And I don't know how much of that is just the biases of the world I've grown up in and what I have been taught and probably more than I realize. But I think more individualism, more human will, more ability to self-determine is important. And also, I think the ability to try new things and not need permission and not need some sort of central planning, betting on human ingenuity and this sort of distributed process, I believe is always going to beat centralized planning. And I think that for all of the deep flaws of America, I think it is the greatest place in the world" }, { "from": "human", "value": "because it's the best at this. So, it's really interesting that centralized planning failed in such big ways. But what if, hypothetically the centralized planning... It was perfect, super intelligent AGI. Super intelligent AGI. Again, it might go wrong in the same kind of ways but it might not and we don't really know." }, { "from": "gpt", "value": "We don't really know. It might be better. it would be better, but would it be better than a hundred super-intelligent or a thousand super-intelligent AGIs sort of in a liberal democratic system? Arguably. Yes. Now, also how much of that can happen internally in one super-intelligent AGI?" }, { "from": "human", "value": "Not so obvious. There is something about, right," }, { "from": "gpt", "value": "but there is something about like tension, the competition." }, { "from": "human", "value": "But you don't know that's not happening inside one model. Yeah, that's true. It'd be nice if, whether it's engineered in or revealed to be happening, it'd be nice for it to be happening." }, { "from": "gpt", "value": "Yeah, of course it can happen" }, { "from": "human", "value": "with multiple AGIs talking to each other or whatever. There's something also about, Mr. Russell has talked about the control problem of always having AGI to have some degree of uncertainty, not having a dogmatic certainty to it. That feels important. So some of that is already handled with human alignment, human feedback, reinforcement learning with human feedback, but it feels like there has to be engineered in like a hard uncertainty. Humility, you can put a romantic word to it. Yeah." }, { "from": "gpt", "value": "Do you think that's possible to do? The definition of those words, I think, the details really matter, but as I understand them, yes, I do. What about the off switch? That like big red button in the data center, we don't tell anybody about it." }, { "from": "human", "value": "I'm a fan, my backpack. I'm getting your backpack. Do you think that's possible to have a switch? You think, I mean, actually more seriously, more specifically about sort of rolling out of different systems. Do you think it's possible to roll them, unroll them," }, { "from": "gpt", "value": "pull them back in? Yeah, I mean, we can absolutely take a model back off the internet." }, { "from": "human", "value": "We can like take, we can turn an API off. Isn't that something you worry about? Like when you release it and millions of people are using it, and like you realize, holy crap, they're using it for, I don't know," }, { "from": "gpt", "value": "worrying about the like all kinds of terrible use cases. We do worry about that a lot. I mean, we try to figure out what this much red teaming and testing ahead of time as we do, how to avoid a lot of those, but I can't emphasize enough how much the collective intelligence and creativity of the world will beat OpenAI and all of the red teamers we can hire. So we put it out, but we put it out in a way we can make changes." } ]

[ { "from": "human", "value": "In the millions of people that've used the chat GPT and GPT, what have you learned about human civilization in general? I mean, the question I ask is, are we mostly good or is there a lot of malevolence in the human spirit?" }, { "from": "gpt", "value": "Well, to be clear, I don't notice anyone else at OpenAI that they're like reading all the chat GPT messages, but from what I hear people using it for, at least the people I talk to, and from what I see on Twitter, we are definitely mostly good, but A, not all of us are, all the time, and B, we really wanna push on the edges of these systems. And, you know, we really wanna test out" }, { "from": "human", "value": "some darker theories for the world, for the world. Yeah, it's very interesting, very interesting. And I think that's not, that actually doesn't communicate the fact that we're like fundamentally dark inside, but we like to go to the dark places in order to maybe rediscover the light. It feels like dark humor is a part of that. Some of the darkest, some of the toughest things you go through if you suffer in life in a war zone, the people I've interacted with that are in the midst of a war, they're usually joking around, joking around, and they're dark jokes. Yeah. So that there's something there. I totally agree about that tension. So just to the model, how do you decide what is and isn't misinformation? How do you decide what is true? You actually have OpenAI's internal factual performance benchmark. There's a lot of cool benchmarks here." }, { "from": "gpt", "value": "How do you build a benchmark for what is true? You should be checking around. There's still many jokes." }, { "from": "human", "value": "Yeah, there's something there." }, { "from": "gpt", "value": "What is truth, Sam Albin? Like math is true, and the origin of COVID is not agreed upon as ground truth. Those are the two things. And then there's stuff that's like certainly not true, but between that first and second milestone," }, { "from": "human", "value": "there's a lot of disagreement. What do you look for? What can a, not even just now, but in the future, where can we as a human civilization look for," }, { "from": "gpt", "value": "look to for truth? What do you know is true?" }, { "from": "human", "value": "What are you absolutely certain is true? I have generally epistemic humiliated about everything and I'm freaked out by how little I know and understand about the world, so that even that question is terrifying to me. There's a bucket of things that have a high degree of truthiness," }, { "from": "gpt", "value": "which is where you put math, a lot of math. Can't be certain, but it's good enough" }, { "from": "human", "value": "for this conversation where you can say math is true. Yeah, I mean, some, quite a bit of physics. There's historical facts, maybe dates of when a war started. There's a lot of details about military conflict inside history. Of course, you start to get just read blitzed, which is this. Oh, I wanna read that." }, { "from": "gpt", "value": "Yeah. How was it? Oh, I wanna read that." }, { "from": "human", "value": "Yeah. How was it? It was really good. It gives a theory of Nazi Germany and Hitler that so much can be described about Hitler and a lot of the upper echelon of Nazi Germany through the excessive use of drugs. Just amphetamines, right? And amphetamines, but also other stuff, but it's just not a lot. And that's really interesting. It's really compelling. And for some reason, whoa, that's really, that would explain a lot. That's somehow really sticky. It's an idea that's sticky. I think you'll really read a lot of criticism of that book later by historians, but that's actually, there's a lot of cherry picking going on. And it actually is using the fact that that's a very sticky explanation. There's something about humans that likes a very simple narrative. Far sure." }, { "from": "gpt", "value": "For sure. For sure. Just amphetamines, right? Or they were already described. Because the war is like a great, even if not true, simple explanation that feels satisfying and excuses a lot of other probably much darker human truths." }, { "from": "human", "value": "Yeah, the military strategy employed the atrocities, the speeches, just the way Hitler was as a human being, the way Hitler was as a leader, all that could be explained through this one little lens. It's like, well, if you say that's true, that's a really compelling truth. So maybe truth is in one sense is defined as a thing that is a collective intelligence we kind of all our brains are sticking to. And we're like, yeah, yeah, yeah, yeah, a bunch of ants get together and like, yeah, this is it. I was going to say sheep, but there's a connotation to that. But yeah, it's hard to know what is true." }, { "from": "gpt", "value": "And I think when constructing a GPT-like model, you have to contend with that. I think a lot of the answers, you know, like if you ask GPT-4, I don't know, just to stick on the same topic, did COVID leak from a lab. I expect you would get a reasonable answer." }, { "from": "human", "value": "There's a really good answer, yeah. It laid out the hypotheses. The interesting thing it said, which is refreshing to hear, is there's something like there's very little evidence for either hypothesis, direct evidence, which is important to state. A lot of people kind of... the reason why there's a lot of uncertainty and a lot of" }, { "from": "gpt", "value": "debate is because there's not strong physical evidence of either. Heavy circumstantial evidence on either side." }, { "from": "human", "value": "And then the other is more like biological theoretical kind of discussion. And I think the answer, the nuanced answer, to the GPT provider was actually pretty damn good. And also importantly, saying that there is uncertainty." }, { "from": "gpt", "value": "Just the fact that there is uncertainty is a statement that was really powerful. And remember when like the social media platforms" } ]

[ { "from": "human", "value": "were banning people for saying it was a lab leak? Yeah, that's really humbling. The humbling, the overreach of power in censorship, but the more powerful GPT becomes," }, { "from": "gpt", "value": "the more pressure there'll be to censor. We have a different set of challenges faced by the previous generation of companies, which is people talk about free speech issues with GPT, but it's not quite the same thing. It's not like, this is a computer program and it's allowed to say, and it's also not about the mass spread and the challenges that I think may have made the Twitter and Facebook and others have struggled with so much. So we will have very significant challenges," }, { "from": "human", "value": "but they'll be very new and very different. And maybe, yeah, very new, very different is a good way to put it. There could be truths that are harmful and they're truth. I don't know. Group differences in IQ. There you go. Yeah, scientific work that when spoken might do more harm. And you ask GPT that, should GPT tell you? There's books written on this that are rigorous scientifically, but are very uncomfortable and probably not productive in any sense, but maybe are. There's people arguing all kinds of sides of this and a lot of them have hate in their heart. And so what do you do with that? If there's a large number of people who hate others but are actually citing scientific studies, what do you do with that? What does GPT do with that? What is the priority of GPT to decrease the amount of hate in the world?" }, { "from": "gpt", "value": "Is it up to GPT or is it up to us humans? I think we as open AI have responsibility for the tools we put out into the world. I think the tools themselves can't have responsibility in the way I understand it." }, { "from": "human", "value": "Wow, so you carry some of that burden." }, { "from": "gpt", "value": "For sure, all of us." }, { "from": "human", "value": "All of us at the company. So there could be harm caused by this tool." }, { "from": "gpt", "value": "There will be harm caused by this tool. There will be tremendous benefits, but tools do wonderful good and real bad and we will minimize the bad and maximize the good." }, { "from": "human", "value": "And you have to carry the weight of that. How do you avoid GPT for being hacked or jailbroken? There's a lot of interesting ways that people have done that, like with token smuggling or other methods like Dan." }, { "from": "gpt", "value": "You know, when I was like a kid, basically I got worked once on jailbreaking an iPhone, the first iPhone, I think. And I thought it was so cool. And I will say it's very strange" }, { "from": "human", "value": "to be on the other side of that. You're now the man. Kind of sucks. Is that, is some of it fun? How much of it is a security threat? I mean, what, how much do you have to take seriously? How is it even possible to solve this problem? Where does it rank on the set of problems? I was just keeping asking questions, prompting." }, { "from": "gpt", "value": "We want users to have a lot of control and get the model to behave in the way they want within some very broad bounds. And I think the whole reason for jailbreaking is right now we haven't yet figured out how to like give that to people. And the more we solve that problem," }, { "from": "human", "value": "I think the less need there will be for jailbreaking." }, { "from": "gpt", "value": "Yeah, it's kind of like piracy gave birth to Spotify. People don't really jailbreak iPhones that much anymore. And it's gotten harder for sure, but also like you can just do a lot of stuff now." }, { "from": "human", "value": "Just like with jailbreaking, I mean, there's a lot of hilarity that is in. So Evan Murakawa, cool guy, he's at OpenAI. He tweeted something that he also was really kind to send me. To communicate with me, sent me a long email describing the history of OpenAI, all the different developments. He really lays it out. I mean, that's a much longer conversation of all the awesome stuff that happened. It's just amazing. But his tweet was Dolly, July 22, Chad GPT, November 22, API 66% cheaper, August 22, embeddings 500 times cheaper while state of the art, December 22, Chad GPT API also 10 times cheaper while state of the art, March 23, Whisper API, March 23, GPT 4 today, whenever that was last week. And the conclusion is this team ships. We do. What's the process of going, and then we can extend that back. I mean, listen, from the 2015 OpenAI launch, GPT, GPT 2, GPT 3, OpenAI 5 finals with the gaming stuff, which is incredible, GPT 3 API released, Dolly, instruct GPT tech, I could find fine tuning. There's just a million things available, the Dolly, Dolly 2 preview, and then Dolly is available to one million people, Whisper, a second model release, just across all of this stuff, both research and deployment of actual products that could be in the hands of people. What is the process of going from idea to deployment that allows you to be so successful" }, { "from": "gpt", "value": "at shipping AI-based products? I mean, there's a question of should we be really proud of that or should other companies be really embarrassed? And we believe in a very high bar for the people on the team. We work hard, which you're not even supposed to say anymore or something. We give a huge amount of trust and autonomy and authority to individual people, and we try to hold each other to very high standards. And there's a process which we can talk about, but it won't be that illuminating. I think it's those other things" } ]

[ { "from": "human", "value": "that make us able to ship at a high velocity. So GPT-4 is a pretty complex system. Like you said, there's like a million little hacks you can do to keep improving it. There's the cleaning up the data set, all that. All those are like separate teams. So do you give autonomy? Is there just autonomy to these fascinating" }, { "from": "gpt", "value": "and different problems? If like most people in the company weren't really excited to work super hard and collaborate well on GPT-4 and thought other stuff was more important, there'd be very little AI or anybody else could do to make it happen. But we spend a lot of time figuring out what to do, getting on the same page about why we're doing something," }, { "from": "human", "value": "and then how to divide it up and all coordinate together. So then you have like a passion for the goal here. So everybody's really passionate across the different teams. Yeah, we care. How do you hire? How do you hire great teams? The folks I've interacted with open AI" }, { "from": "gpt", "value": "are some of the most amazing folks I've ever met. It takes a lot of time. Like I spend, I mean, I think a lot of people claim to spend a third of their time hiring. I for real truly do. I still approve every single hired open AI. And I think there's, you know, we're working on a problem that is like very cool and the great people want to work on. We have great people and some people want to be around them. But even with that, I think there's just no shortcut" }, { "from": "human", "value": "for putting a ton of effort into this." }, { "from": "gpt", "value": "So even when you have the good people hard work." }, { "from": "human", "value": "I think so. Microsoft announced the new multi-year, multi-billion dollar reported to be $10 billion investment into open AI. Can you describe the thinking that went into this? And what are the pros, what are the cons" }, { "from": "gpt", "value": "of working with a company like Microsoft? It's not all perfect or easy, but on the whole, they have been an amazing partner to us. Satya and Kevin and Mikael are super aligned with us, super flexible, have gone like way above and beyond the call of duty to do things that we have needed to get all this to work. This is like a big iron complicated engineering project. And they are a big and complex company. And I think like many great partnerships or relationships, we've sort of just continued to ramp up our investment in each other." }, { "from": "human", "value": "And it's been very good. It's a for-profit company. It's very driven. It's very large scale." }, { "from": "gpt", "value": "Is there pressure to kind of make a lot of money? And I think most other companies wouldn't, maybe now they would, it wouldn't at the time have understood why we needed all the weird controller provisions we have and why we need all the kind of like AGI specialness. And I know that cause I talked to some other companies before we did the first deal with Microsoft. I think they were, they are unique in terms of the companies at that scale that understood why we needed the control provisions we have." }, { "from": "human", "value": "So those control provisions help you, help make sure that the capitalist imperative does not affect the development of AI. Well, let me just ask you as an aside about Sachin Adela, the CEO of Microsoft. He seems to have successfully transformed Microsoft into this fresh, innovative, developer-friendly company. I agree. What do you, I mean, it's really hard to do for a very large company. What have you learned from him? Why do you think he was able to do this kind of thing? Yeah, what insights do you have about why this one human being is able to contribute to the pivot of a large company into something very new?" }, { "from": "gpt", "value": "I think most CEOs are either great leaders or great managers. And from what I have observed with Satya, he is both. Supervisionary really gets people excited, really makes long duration and correct calls. And also he is just a super effective hands-on executive and I assume manager too." }, { "from": "human", "value": "And I think that's pretty rare. I mean, Microsoft, I'm guessing like IBM or like a lot of companies have been at it for a while, probably have like old school kind of momentum. So you like inject AI into it, it's very tough, right? Or anything, even like open source, the culture of open source. Like how hard is it to walk into a room and be like, the way we've been doing things are totally wrong. Like I'm sure there's a lot of firing involved or a little like twisting of arms or something. So do you have to rule by fear, by love?" }, { "from": "gpt", "value": "Like what can you say to the leadership aspect of this? I mean, he's just like done an unbelievable job, but he is amazing at being like clear and firm and getting people to want to come along." }, { "from": "human", "value": "But also like compassionate and patient with his people too. I'm getting a lot of love, not fear. I'm a big Satya fan. So am I from a distance. I mean, you have so much in your life trajectory that I can ask you about, we can probably talk for many more hours, but I got to ask you because of Y Combinator, because of startups and so on. The recent, you've tweeted about this, about the Silicon Valley bank, SVB. What's your best understanding of what happened? What is interesting? What is interesting to understand" }, { "from": "gpt", "value": "about what happened with SVB? I think they just like horribly mismanaged buying while chasing returns in a very silly world of 0% interest rates. Buying very long dated instruments secured by very short term and variable deposits. And this was obviously dumb. I think totally the fault of the management team, although I'm not sure what the regulators were thinking either. And is an example of where I think you see the dangers of incentive misalignment, because as the Fed kept raising, I assume that the incentives on people working at SVB to not sell at a loss, they're super safe bonds, which were now down 20% or whatever, or down less than that, but then kept going down. That's like a classy example of incentive misalignment. Now, I suspect they're not the only bank in the bad position here. The response of the federal government, I think took much longer than it should have, but by Sunday afternoon, I was glad they had done what they've done." } ]

[ { "from": "human", "value": "We'll see what happens next. So how do you avoid depositors" }, { "from": "gpt", "value": "from doubting their bank bank? What I think needs would be good to do right now is just, and this requires statutory change, but it may be a full guarantee of deposits, maybe a much, much higher than 250K, but you really don't want depositors having to doubt the security of their deposits. And this thing that a lot of people on Twitter were saying is like, well, it's their fault. They should have been reading the balance sheet and the risk audit of the bank. Do we really want people to have to do that?" }, { "from": "human", "value": "I would argue no." }, { "from": "gpt", "value": "What impact has it had on startups that you see? Well, there was a weekend of terror, for sure. And now I think, even though it was only 10 days ago," }, { "from": "human", "value": "it feels like forever and people have forgotten about it." }, { "from": "gpt", "value": "But it kind of reveals the fragility of our economic system. We may not be done. That may have been like the gun show and falling off the nightstand in the first scene of the movie or whatever. It could be like other banks." }, { "from": "human", "value": "For sure there could be, for sure there could be. Well, even with FTX, I mean, I'm just, well, that's fraud, but there's mismanagement. And you wonder how stable our economic system is, especially with new entrants with AGI." }, { "from": "gpt", "value": "I think one of the many lessons to take away from this SVP thing is how much, how fast and how much the world changes and how little, I think, are experts, leaders, business leaders, regulators, whatever, understand it. So the speed with which the SVP bank run happened because of Twitter, because of mobile banking apps, whatever, was so different than the 2008 collapse where we didn't have those things really. And I don't think that kind of the people in power realize how much the field had shifted. And I think that is a very tiny preview" }, { "from": "human", "value": "of the shifts that AGI will bring. What gives you hope in that shift from an economic perspective?" }, { "from": "gpt", "value": "Ah, it sounds scary, the instability. No, I am nervous about the speed with which this changes and the speed with which our institutions can adapt, which is part of why we want to start deploying these systems really early, why they're really weak, so that people have as much time as possible to do this. I think it's really scary to have nothing, nothing, nothing and then drop a super powerful AGI all at once on the world. I don't think people should want that to happen. But what gives me hope is I think the less zeros, the more positive some of the world gets, the better. And the upside of the vision here, just how much better life can be. I think that's gonna like unite a lot of us and even if it doesn't, it's just gonna make it all feel more positive some." }, { "from": "human", "value": "When you create an AGI system, you'll be one of the few people in the room that get to interact with it first, assuming GPT-4 is not that. What question would you ask her, him, it?" }, { "from": "gpt", "value": "What discussion would you have? One of the things that I have realized, this is a little aside and not that important, but I have never felt any pronoun other than it towards any of our systems. But most other people say him or her or something like that. And I wonder why I am so different. Yeah, I don't know, maybe it's I watch it develop, maybe it's I think more about it," }, { "from": "human", "value": "but I'm curious where that difference comes from. I think probably because you watch it develop, but then again, I watch a lot of stuff develop and I always go to him or her. I anthropomorphize aggressively" }, { "from": "gpt", "value": "and certainly most humans do. I think it's really important that we try to explain," }, { "from": "human", "value": "educate people that this is a tool and not a creature. I think I, yes, but I also think there will be a room in society for creatures" }, { "from": "gpt", "value": "and we should draw hard lines between those. If something's a creature, I'm happy for people to like think of it and talk about it as a creature, but I think it is dangerous to project creatureness" }, { "from": "human", "value": "onto a tool. That's one perspective. A perspective I would take if it's done transparently is projecting creatureness onto a tool" }, { "from": "gpt", "value": "makes that tool more usable if it's done well. Yeah, so if there's like kind of UI affordances that work, I understand that." }, { "from": "human", "value": "I still think we wanna be like pretty careful with it. Because the more creature like it is, the more it can manipulate you emotionally." }, { "from": "gpt", "value": "Or just the more you think that it's doing something or should be able to do something" }, { "from": "human", "value": "or rely on it for something that it's not capable of. What if it is capable? What about Sam Albin? What if it's capable of love? Do you think there will be romantic relationships" }, { "from": "gpt", "value": "like in the movie Her or GPT? There are companies now that offer, like for lack of a better word, like romantic companion ship AIs. Replica is an example of such a company." }, { "from": "human", "value": "Yeah, I personally don't feel any interest in that. So you're focusing on creating intelligent tools. But I understand why other people do. I understand why other people do. That's interesting." }, { "from": "gpt", "value": "I have, for some reason, I'm very drawn to that. It's interesting. Have you spent a lot of time interacting" }, { "from": "human", "value": "with Replica or anything similar? Replica, but also just building stuff myself. Like I have robot dogs now that I use, I use the movement of the robots to communicate emotion." }, { "from": "gpt", "value": "I've been exploring how to do that. Look, there are gonna be very interactive GPT-4 powered pets or whatever, robots, companions, and companions." } ]

[ { "from": "human", "value": "A lot of people seem really excited about that. Yeah, there's a lot of interesting possibilities. I think you'll discover them, I think, as you go along. That's the whole point. Like the things you say in this conversation, you might in a year say this was right." }, { "from": "gpt", "value": "No, I may totally want, I may turn out" }, { "from": "human", "value": "that I love my GPT-4 dog robot or whatever. Maybe you want your programming assistant to be a little kinder and not mock you." }, { "from": "gpt", "value": "I hear incompetence. No, I think you do want the style of the way GPT-4 talks to you really matters. You probably want something different than what I want, but we both probably want something different than the current GPT-4. And that will be really important," }, { "from": "human", "value": "even for a very tool-like thing. Yes. Is there styles of conversation? No, contents of conversations you're looking forward to with an AGI, like GPT-5, 6, 7. Is there stuff where, like where do you go to" }, { "from": "gpt", "value": "outside of the fun meme stuff for actual life? I mean, what I'm excited for is like, please explain to me how all the physics works and solve all remaining mysteries." }, { "from": "human", "value": "So like a theory of everything. I'll be real happy. Faster than light travel. Don't you want to know? So there's several things to know. It's like, and be hard. Is it possible in how to do it? Yeah, I want to know. I want to know. Probably the first question would be, are there other intelligent alien civilizations out there? But I don't think AGI has the ability to do that," }, { "from": "gpt", "value": "to know that. Might be able to help us figure out how to go detect. And we need to like send some emails to humans and say, can you run these experiments? Can you build the space probe? Can you wait a very long time?" }, { "from": "human", "value": "We'll provide a much better estimate than the Drake equation. With the knowledge we already have. And maybe process all the, because we've been collecting a lot of." }, { "from": "gpt", "value": "Yeah, you know, maybe it's in the data. Maybe we need to build better detectors, which in a really advanced way I could tell us how to do. It may not be able to answer it on its own, but it may be able to tell us what to go build, to collect more data." }, { "from": "human", "value": "What if it says the alien?" }, { "from": "gpt", "value": "What if it says the alien's already here?" }, { "from": "human", "value": "I think I would just go about my life." }, { "from": "gpt", "value": "I mean, a version of that is like, what are you doing differently now that like, if GPT-4 told you and you believed it, okay, AGI is here, or AGI is coming real soon." }, { "from": "human", "value": "What are you gonna do differently? The source of joy and happiness and fulfillment of life is from other humans. So it's mostly nothing. Unless it causes some kind of threat." }, { "from": "gpt", "value": "But that threat would have to be like literally a fire. Like are we living now with a greater degree of digital intelligence than you would have expected three years ago in the world? And if you could go back and be told by an Oracle three years ago, which is blink of an eye, that in March of 2023, you will be living with this degree of digital intelligence, would you expect your life to be more different" }, { "from": "human", "value": "than it is right now? Probably, probably. But there's also a lot of different trajectories intermixed. I would have expected the society's response to a pandemic to be much better, much clearer, less divided. I was very confused about. There's a lot of stuff, given the amazing technological advancements that are happening, the weird social divisions, it's almost like the more technological advancement there is, the more we're going to be having fun with social division. Or maybe the technological advancement is just to reveal the division that was already there. But all of that just confuses my understanding of how far along we are as a human civilization and what brings us meaning and how we discover truth together and knowledge and wisdom. So I don't know, but when I open Wikipedia, I'm happy that humans were able to create this thing. For sure. Yes, there is bias, yes. Let's think for a second. It's a triumph. It's a triumph of human civilization. 100%. Google search, the search, search, period, is incredible. The way it was able to do 20 years ago. And now this new thing, GPT, is this gonna be the next, the conglomeration of all of that that made web search and Wikipedia so magical, but now more directly accessible. You're kind of a conversation with a damn thing. It's incredible. It's a triumph. Let me ask you for advice for young people. In high school and college, what to do with their life, how to have a career they can be proud of, how to have a life they can be proud of. You wrote a blog post a few years ago titled How to be Successful. And there's a bunch of really, really, people should check out that blog post. They're so, it's so succinct and so brilliant. You have a bunch of bullet points. Compound yourself, have almost too much self-belief, learn to think independently, get good at sales and quotes, make it easy to take risks, focus, work hard, as we talked about, be bold, be willful, be hard to compete with, build a network. You get rich by owning things, be internally driven. What stands out to you from that" }, { "from": "gpt", "value": "or beyond as advice you can give? Yeah, no, I think it is like good advice in some sense, but I also think it's way too tempting to take advice from other people. And the stuff that worked for me, which I tried to write down there, probably doesn't work that well, or may not work as well for other people. Or like other people may find out that they want to just have a super different life trajectory. And I think I mostly got what I wanted by ignoring advice. And I think like I tell people not to listen to too much advice. Listening to advice from other people should be approached with great caution." } ]

[ { "from": "human", "value": "How would you describe how you've approached life? Outside of this advice, that you would advise to other people. So really just in the quiet of your mind to think what gives me happiness? What is the right thing to do here?" }, { "from": "gpt", "value": "How can I have the most impact? I wish it were that introspective all the time. It's a lot of just like, what will bring me joy? What will bring me fulfillment? I do think a lot about what I can do that will be useful, but like who do I want to spend my time with?" }, { "from": "human", "value": "What I want to spend my time doing?" }, { "from": "gpt", "value": "Like a fish in water is going along with it. Yeah, that's certainly what it feels like. I mean, I think that's what most people would say" }, { "from": "human", "value": "if they were really honest about it. Yeah, if they really think, yeah. And some of that then gets to the Sam Harris discussion of free wellbeing and illusion. Which is very well might be, which is a really complicated thing to wrap your head around. What do you think is the meaning of this whole thing? That's a question you could ask an AGI. What's the meaning of life? As far as you look at it, you're part of a small group of people that are creating something truly special. Something that feels like," }, { "from": "gpt", "value": "almost feels like humanity was always moving towards. Yeah, that's what I was gonna say is, I don't think it's a small group of people. I think this is the, I think this is like the product of the culmination of whatever you wanna call it, an amazing amount of human effort. And if you think about everything that had to come together for this to happen, when those people discovered the transistor in the 40s, like, is this what they were planning on? All of the work, the hundreds of thousands, millions of people, whatever, it's been that it took to go from that one first transistor to packing the numbers we do into a chip and figuring out how to wire them all up together. And everything else that goes into this, you know, the energy required, the science, like just every, every step. Like, this is the output of like all of us." }, { "from": "human", "value": "And I think that's pretty cool. And before the transistor, there was a hundred billion people who lived and died, had sex, fell in love, ate a lot of good food, murdered each other sometimes, rarely, but mostly just good to each other, struggled to survive. And before that, there was bacteria and eukaryotes and all that. And all of that was on this one exponential curve. Yeah, how many others are there? I wonder. We will ask, that isn't question number one for me, for AGI, how many others? And I'm not sure which answer I want to hear. Sam, you're an incredible person. It's an honor to talk to you. Thank you for the work you're doing. Like I said, I've talked to Ilya Seskera, I talked to Greg, I talked to so many people at OpenAI. They're really good people." }, { "from": "gpt", "value": "They're doing really interesting work. We are gonna try our hardest to get to get to a good place here. I think the challenges are tough. I understand that not everyone agrees with our approach of iterative deployment and also iterative discovery, but it's what we believe in. I think we're making good progress and I think the pace is fast, but so is the progress. So like the pace of capabilities and change is fast, but I think that also means we will have new tools to figure out alignment" }, { "from": "human", "value": "and sort of the capital S safety problem. I feel like we're in this together. I can't wait what we together" }, { "from": "gpt", "value": "as a human civilization come up with. It's gonna be great, I think." }, { "from": "human", "value": "We'll work really hard to make sure. Me too, me too. Thanks for listening to this conversation with Sam Altman. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words from Alan Turing in 1951. It seems probable that once the machine thinking method has started, it would not take long to outstrip our feeble powers. At some stage, therefore, we should have to expect the machines to take control. Thank you for listening and hope to see you next time." } ]

[ { "from": "human", "value": "Bukowski." }, { "from": "gpt", "value": "Does it have to fade? Bukowski. The truth is that you have all of these chemicals pumping through your body. You're essentially high on heroin in the beginning of a romance. And you're going to have these rose colored glasses on everything your partner does is magical. But really, it's the novelty. It's just like going on a vacation. You're fully present. You're just attuned to the magic of another human being moment to moment. And then on top of that, you have, you're just flooded with dopamine. So you're high on drugs. And we can't go on like that. You will die if you are using these kinds of chemicals all the time, all day long. So eventually, our bodies are sort of made to dial it down. We've made it, I mean, we're evolutionary beings. We are doing the same thing we did 200,000 years ago to find a mate, procreate, spend enough time with each other that we have sex a whole bunch of times and make babies. Now we've changed the rules of the game. We're living almost till we're 100 years old in some cases. We're making these marriage commitments that last half a century. And we're expecting it to be all because of love. And we're signing these contracts based on how we feel when we're high on these drugs. So the reality is we know based on the, and I'm also talking about certain Western civilizations here because as you know, there are arranged marriages. And a lot of times those marriages, if we're looking at longevity, are actually way more satisfied than people who are marrying for love, which logically makes sense. If you're making a decision based on a feeling that is basically based on endorphins and dopamine and oxytocin, I wouldn't sign a contract just because of a feeling necessary for 50 years. Whereas an arranged marriage, if you have your elders kind of deciding for you that this partner has a bunch of traits that you're going to appreciate more and more over time," }, { "from": "human", "value": "I think there's some wisdom there. So you don't think that feeling could be a foundation" }, { "from": "gpt", "value": "for a 50 year relationship? Well, I don't think that specific feeling you're having based on drugs is going to be the same feeling you have 20, 30, 40 years down the line. If you're going to wake up and turn to your partner when you're 70 and think, oh my God, I'm so glad you're hot, you are so hot, then sure marry for hotness. But if you've been through life a little bit, and I think most people who are on a second marriage know, shit happens in life. It is hard, you're going to have, maybe a kid with special needs or your dad gets dementia or you get diagnosed with cancer. Who are you going to want to come home to? Who is going to hold you when you are sobbing on the floor and tell you we're going to get through it together? Who's going to know the names of your kid's special ed teacher and the process for getting a 504 plan? Or is it going to be you on your own?" }, { "from": "human", "value": "I think those things matter. Yeah, but doesn't that hotness, don't those drugs kind of solidify into a deeper appreciation of the other person, into something you could call beauty?" }, { "from": "gpt", "value": "Yes, they can." }, { "from": "human", "value": "But isn't that the same thing? When you notice the beauty of another human being, aren't you high on drugs still? You're making it sound like there's a brief rock star period of going on heroin and then it's over, but can't you be on heroin your whole life? I have some good news." }, { "from": "gpt", "value": "Is that the same thing? I have some good news. That was something, I think one of the reasons I got into studying relationships was because I wanted that, right? So I'm a scientist, but I also love art and I love writing and I love literature. I wanted to know that true love could be real, but as a scientist, I am cynical. I just need some data. And so I practice a type of therapy called the Gottman method. And I love that because it tends to be, well, it is one of the most evidence-based therapies we have based on John and Julie Gottman, two psychologists who have been researching relationships for now about 50 years. And this therapy happens to be for couples. They found that you absolutely can make longevity work in a relationship. You can build, you are not just settling for companionship, but you can have passion and intimacy and growing love and appreciation. But there is a blueprint, a set of skills that we were never given. We're not taught it in school. We changed the rules of the game" }, { "from": "human", "value": "and we haven't learned the rules yet. And the Gottman method for couples therapy kind of gives you a few guidelines," }, { "from": "gpt", "value": "the rules for longevity in a relationship. For longevity in a relationship. Yeah, they did a beautiful job at taking these findings they had through decades of research, quantifying it, and then codifying it into a therapy method. It's really skills-based. I tell couples when they're starting out with me" }, { "from": "human", "value": "that they're essentially gonna be starting a class. So what's the five to one golden rule? What I read is just the kind of balance you can achieve of how many interactions you have in relationships that are positive versus negative. And I think that's what the five to one means, but basically there should be kind of an empirical rule. Like if you just look back over a month, how many of the interactions were positive, how many were negative? Or the day." }, { "from": "gpt", "value": "Or the day. Or the day. The day, right? So the idea of this ratio, well, it's not an idea. It was a finding. It is a research finding that the Gottman's got after looking at thousands of couples and codifying these interactions that they were observing. Couples that tend to be satisfied in their relationships that are happier, they have better health, et cetera, they are having approximately five positive interactions to each negative. And I wanna be clear about what I'm defining as positive and negative here. So this doesn't necessarily mean that you're, these don't need to be big sweeping romantic gestures, buying flowers, having sex. These are things like paying attention to what we call your partner's bids. We make these bids for affection, for connection, all the time in our relationships, not just with our partners, but with our friends, our coworkers. And we may not even know what our style of bid is, but if you see them on a sheet, you can pretty quickly identify them. Bids could be wanting to show your partner, tell your partner something and have them be proud of you. It could be wanting to go buy groceries with your partner doing things together. Hey, you wanna come with me? It could be telling a joke and hoping that your wife looks up from her email on the computer and acknowledges it. If she laughs, then you've got a positive. But if I don't even look up, that's a negative, right? So it's not necessarily that I'm calling my husband an asshole, it's just, am I connecting with him?" } ]

[ { "from": "human", "value": "Am I meeting those bids for connection and vice versa? Well, do those also give you a guide" }, { "from": "gpt", "value": "of how you should behave? Well, I think what's really important is actually asking your partner or paying attention to what your partner's bids are, because what matters to Ty, my husband, may not matter to you. For instance, Ty's bar is so low with me." }, { "from": "human", "value": "Thank God for him, thank God for him." }, { "from": "gpt", "value": "What defines a positive interaction? Right, like he just wants me to ask him if he wants a water when I get up to get myself one. Just be a basic, decent, considerate person is all he asks of me. Whereas mine might be sort of like, stay up later with me, watch a show, go to bed at the same time as me" }, { "from": "human", "value": "or know about the people in my life, that sort of a thing. I should highlight this and I hope, hopefully it's okay, that you were running a little bit late and you sent me this text, which people do really rarely, and there's a subtle act of kindness within that text. So the text you sent was that I just decreased the amount of stress in your life or something like this by saying it's cool. But that means that you're signaling that you were stressed because you care enough to be there on time. And that made me feel really special, like, oh, people don't often, people don't always do that because that puts you also, that makes you vulnerable. Vulnerable." }, { "from": "gpt", "value": "And that actually is by saying, people don't know. You're a brilliant guy, vulnerable. I actually thought that after I sent it, but I feel that most of the day, any interaction, like, oh God, I just exposed myself. But absolutely, I was excited to be here and I didn't want you to think that I didn't care. I think being a therapist has shown me that it really, it's so lucky to be in that position because you meet people that you would have thought are cooler than you or smarter than you or just somehow impervious to life. And you realize that we are all in it together. We all want to be cared about and liked. We don't want to be liked as a baseline. Some people will say they don't care, but everybody does. It's human. And I have gotten much better being a therapist, much more comfortable showing caring, showing love and genuineness and vulnerability" }, { "from": "human", "value": "than I think I ever would have been otherwise. Any interaction like that. And that kind of vulnerability is what's required" }, { "from": "gpt", "value": "to do a positive interaction in a relationship? I think so. And people have different levels of comfort, right? So, but as long as it's working for both partners and typically you have to communicate to figure out what your partner, what makes your partner feel cared about. However, you might be working, for instance, with an older couple, and I have a couple that's perfectly happy and they sort of have a system. It works for them. If there's some sort of a rupture, if they get in some sort of a disagreement, they don't talk it out. She might go to the store, run an errand, do a little shopping. He'll work in the wood shop. And then they'll come back and there is a repair attempt though, but it's maybe she'll say, hey, do you want to have dinner? Come, you know, I made your favorite dinner. Or he'll say, hey, I recorded your favorite show. You want to watch it tonight? So they don't need to process it, but there is an understanding between them that we're still in this together, we care about each other, and there's a repair attempt. Most people need to be able to process it verbally" }, { "from": "human", "value": "and talk about what happened, but not all. So for most people, if there's a conflict, you should talk about and resolve it and repair it" }, { "from": "gpt", "value": "versus like just put it behind you? I don't always say should, I guess it depends on the couple. Everybody processes emotions differently, everybody handles emotional expression differently. I mean, I have couples where I have one person in the partnership who has autism and the other doesn't. And so they're obviously going to have different ways of communicating or processing what happened. We all have different perspectives. It really depends on what makes a person feel like it's been repaired. What makes a person feel understood? Does that need to be verbal? Or in the case of that older couple I have where they know they understand one another" }, { "from": "human", "value": "because there's a gentleness toward one another after. What are some common ways relationships fail" }, { "from": "gpt", "value": "that you've observed in all the therapy you've had? Well, the Gomins identified what they call the four horseman of the apocalypse. So, I mean, I could just keep it simple and go off their research. So those are four different behaviors that they identify in couples or that you can identify in couples that are really highly predictive of a divorce, some more than others, but I'll start with the lower ones. So one thing that we, by the way, actually, we all do these things. These would be in that five to one ratio. You'd wanna stay away from some of these. These are the ones. So as they pile up, now that ratio is gonna get imbalanced and then you are headed for a split, okay? So the first is criticism. So criticism is when we have a complaint, complaints are normal, but instead of owning our own problems, our own feelings, we assume that our perspective is the only valid, accurate perspective. And so we take it upon ourselves to tell our partner what is wrong with them, okay?" }, { "from": "human", "value": "So there's essentially no real belief" }, { "from": "gpt", "value": "that they might have a valid perspective too. So this could look like you never helped me out in that with the house or even you're so lazy, like, can't I just get you for five seconds to help with the kids or something like that? And then what happens is horseman number two, defensiveness. So not everybody is defensive just because they were criticized. Some people just are more prone to defensiveness than others. None of us really like admitting our faults. So it's pretty natural, but defensiveness is essentially making excuses or worse, turning it around on your partner, not accepting any responsibility and definitely not validating what they're feeling. Now, if you get criticized enough, or if you get really flooded, the flooding is what happens when our heart rate goes up kind of around a hundred beats per minute, our frontal lobe shuts down. That's our thoughtful brain, our logical brain. And our reptilian kind of hind brain takes over our thinking. And we just go into fight or flight in a way. We just want to annihilate our partner and instead of say anything, that would be helpful to the relationship. So if you're getting flooded, you could do a couple of things. You could get super critical, you could get contemptuous, which I'll talk about in a second, it's the last horseman, or you do the third horseman, which is stonewalling. And in their research, the Gottmans found that men are actually more likely to stonewall. I also am someone who stonewalls, but it's where you just sort of disconnect from the conversation. You shut down, you turn away. You can physically even turn away kind of arms crossed, but you're just, you're shut off. And stonewalling happens usually because you get flooded. You feel like you can't win. You don't know what to do to make the situation better. It feels pretty hopeless and talking feels unproductive. So you can see how in a typical heterosexual relationship, the gender dynamic, we know that women tend to use criticism more often because they're the ones that typically raise issues verbally. And then if men are feeling more criticized, that they tend to stonewall and it becomes this vicious cycle of then more criticism, but the criticism is really just a plea to be loved and get your partner to show you they care. And then the man tends to feel like he can't do anything right. This isn't even productive if I say anything, I'm just gonna make it worse. And they don't have any real, you haven't given them a specific need, a solution, something they can do to shine for you." } ]

[ { "from": "human", "value": "So they turn away. And where does the contempt come in?" }, { "from": "gpt", "value": "All right, all right. So contempt is criticism on steroids. This is what John Gottman calls sulfuric acid for love. Nothing will erode a relationship quicker than contempt. Contempt is when you are looking at your partner from a superior position. So you are eye rolling, you are name calling, there's a mockery, mocking even physical mockery, imitating them, imitating their voice. Contempt is meant to just take the legs out from your partner, make them feel pathetic, ridiculous. And it can be abusive, but most people have engaged in contempt at some point in their relationship, lower level would be sort of the eye rolling," }, { "from": "human", "value": "but that is the biggest predictor of a split. If you allow yourself to think, yeah, that mockery or contempt just a little bit, it's like this weird slippery slope. Sure is. And the opposite is true, where I just look at a person and think, wow, isn't that the most wonderful creature I've ever seen in my life? Just think that. And you notice the little details about who they are. That's why I just observe them the way you observe a weird peacock at a zoo or something like that." }, { "from": "gpt", "value": "A little detail." }, { "from": "human", "value": "Intention is powerful, isn't it? Yeah, and it changes. You start to notice beautiful things and then let the things that annoy you. Yes." }, { "from": "gpt", "value": "You're exactly right. You're touching on some really important things. So in relationships, we actually know that wearing rose-colored glasses is important, it's healthy, we need it. And it's a choice you're making, right? So there is a saying that getting married is just choosing one person's faults over another. And the reality is that we may become infatuated with somebody else as human beings. Love is an emotion, attraction is an emotion. And as you go through life, even if you're in a committed relationship, you might see beauty in another. And that person who is novel might seem attractive to you. But if you can remember that they too have a set of problems that you would be marrying, it really helps you to see the beauty in your partner again and recognize all of their incredible strengths and all the ways we meld with a person and become our own family. Almost become, I mean, our lives intertwine" }, { "from": "human", "value": "and we grow those oak trees, some really important. So by the way, it's a line I read somewhere that when you're wearing rose-colored glasses, all the red flags looks just like flags. I think it's a good line. So you think that humans are fundamentally, all of us are fundamentally flawed or have flaws, they're unique flaws. And they're basically relationships," }, { "from": "gpt", "value": "it's just a way to figure out how the two can fit together. Right, and we're different. So no matter what, we're going to have differences. We are raised differently than our partner. We have different stories, different experiences that shaped our value systems, especially when it comes to the big ones like parenting, love, money, these principles that are based in our history. We're gonna have differences. So is this a set of differences you can accept from somebody and work with? Do the benefits and do their strengths, do they make it worth it?" }, { "from": "human", "value": "Or are they deal breaker differences? Tricky question, but in the couples you've worked with, is there like the feminine and the masculine, is there different dynamics that come into play, like dominant, submissive? Is it like a dance where it just changes from minute to minute? Is there dynamics that you observe that both limit and enable successful relationships?" }, { "from": "gpt", "value": "Yes, so there are, if we're talking about masculine, feminine, then how also are we could get into, are we talking about actual gender, identified gender, or are we just talking about these traits? Because like I said, I stonewall, which is typically in couples, something that is more associated with straight men. But that's my style of coping when I get overwhelmed. That is not tied to any sort of success or non-success of a relationship. But what we do know is that gay couples, so lesbians and gay men, tend to be gentler with one another when they are having conflict discussions. I, so that's actually been identified in the research and it's something I've witnessed and it's just fascinating. So with my straight couples, I'll be going through one of these, if we're processing the conflict that occurred, I'll be going through the sheet and it's very, very structured because you don't want couples doing more damage when they're there with you. You want them practicing skills that protect them from criticism, that protect them from contempt. And when I'm working with a straight couple, I am like a referee or sometimes I'll relate it to being like a ski coach and keeping people on a funny hill and you tell them, you let them make like two turns and then you stop them and you meet up again because you don't want them to veer off. With straight couples, you are doing very short turns before you need to kind of intervene and re-scaffold. I had a lesbian couple recently and they were so lovely with each other. They skipped like seven steps to the advanced final portion where they were already coming up with solutions and suggesting things that they might be able to do differently next time to make it better for their partner. They were asking each other questions about how their partner felt with no agenda, no attempt to sort of be like, well, do you think you're feeling that way because, which straight couples do all the time, you just see this humility and openness." }, { "from": "human", "value": "It's lovely. Yeah, it's lovely. But I wonder if maybe watching too many Hollywood films if some of the drama, some of the tension" }, { "from": "gpt", "value": "is required for a passionate lifelong romance. No, it's not. And that's great news. So we actually know that the closer you feel to your partner. You've talked a lot about beauty and you can ignite that beauty, that interest, right? So when you're falling in love, it's usually that a person is sort of a mystery to you and you're uncovering these layers that you find really appealing. There are continual layers that you can uncover with your partner over time. I don't think we realize that. I think we get complacent and we think we've had every conversation imaginable. What else are they gonna do to surprise me? But we don't know the questions to be asking. One of my favorite questions, I like turning these conversations kind of into a quiz because I get bored easily. So rather than just asking an open-ended question, there's a way you can do this with your partner where it's sort of like the dating game. Like, what is my as of yet fondest but unrealized life dream? And see if your partner knows. You might not even know. They might know you better than you know yourself. That in and of itself is a beautiful reminder of the relationship and how special it is. But then also when they say it or when you realize or have to think critically, like what is my husband's as of yet unrealized but fondest life dream?" } ]

[ { "from": "human", "value": "And then you can talk about it." }, { "from": "gpt", "value": "That you just, I don't know, you just kind of transcend into this new area and you feel tight again." }, { "from": "human", "value": "You feel close. Well, you really talk to each other. Like I've recorded and without intending to publish a podcast like this with microphones, with friends, with people close to me because it's literally that you get to ask questions like as if it's an interview. And we don't do that to each other." }, { "from": "gpt", "value": "The way you're talking with me. That's exactly it. The way you're talking with me. Sit down with your partner," }, { "from": "human", "value": "have that conversation." }, { "from": "gpt", "value": "Like years later. Right, show interest, actually be curious." }, { "from": "human", "value": "See what they surprise you with. And actually what you learn is you don't know the answers to most of these questions. 100%. Exactly. Like what's your favorite movie from the 80s? You might not know the answer to that. It's like those first date questions or whatever." }, { "from": "gpt", "value": "Or 100%. Exactly. Date questions or whatever. Or what's your favorite movie this year and why?" }, { "from": "human", "value": "And why, yeah. It's fascinating. It is. It's hard to do that because I think that you'll probably be offended at first, how little the other person knows." }, { "from": "gpt", "value": "So I think you have to work through that. You know, I actually find that there's this rekindling because partners are shocked that their partner does know so much about them. Especially if they've been feeling dissatisfied or disconnected." }, { "from": "human", "value": "It's a reminder of all the good that's still there. What, I know we've said some of those things, but what's on the opposite side? What's the key to a successful relationship? What's like, what are the things you see time and time again" }, { "from": "gpt", "value": "that to you designate that they're on a good path? Yeah. There's a real attunement, honestly. Just it's sort of an us against the world feeling. Nobody, neither partner is gonna talk shit on them. The other, there's a loyalty. They handle each other in the relationship with care. You can tell that they've worked some things. To me, it usually indicates that these are some people who have figured they've had to work some things out. They know that this is delicate. They know, you know, that you're on thin ice. You take a wrong step and you can be back in a tough place in your relationship or you treat it with care and it can be amazing. So they're careful with one another. They give each other compliments. They are considerate. So you'll see, you know, he'll bring the car around for her because it's raining or she'll bring him home some takeout, you know, she'll order for him too at the restaurant." }, { "from": "human", "value": "There's just, they keep each other in each other's minds. But that us against the world thing, that definitely is there. Like if you- 100%, you've seen that, right? Yeah, you've seen it. And you've seen it like, I like it when couples have been together for a long time and when one is talking, the other one looks at them. If you don't do that, that's not a bad sign, but it's a good sign when you do that. Yes. And I think it's actually a really good exercise to do because I enjoy when I see in others." }, { "from": "gpt", "value": "So it's a- 100%, you've seen that, right?" }, { "from": "human", "value": "Yes. It's a way to show that you don't take him for granted and then you still find them like this mysterious, wonderful creature to observe. Like I think too often we have that with our parents, we have that with people close to us, you think, yeah, I've heard what they're about to say. I know, I know you can complete their sentences. Take him for granted. And then if you just look at them and say, wow, this is the most brilliant person I've ever seen of my life. I can't just appreciate every word that comes out of them and look at them in that way. You actually begin to believe it. And you actually begin to see" }, { "from": "gpt", "value": "the beauty of what they're saying. Take him for granted. You are exactly right. It's a self-fulfilling prophecy. And they go really good. And it's caring, yeah. It's very caring, so that's, I mean, that's I think the beauty of what the Gottman research showed us, taught us, provided us, is that we can do these things that become cyclic and just keep growing this relationship, making it stronger, more powerful, more loving." }, { "from": "human", "value": "You would never want to cut it down. And it's caring, yeah. Well, you were talking about the sheet for conflict processing. What are we talking about? So like a couple will come and say like, there was this conflict and you put it on the table" }, { "from": "gpt", "value": "and then what does it mean to process it? So in that Gottman Methodotherapy, there are all these different, I mean, hundreds of different interventions and based on what the issue is in that session, you can decide the most appropriate intervention. And so this is a specific intervention for if it is a conflict that occurred and there are different types of conflicts. So this would be more like an incident. It's not a perpetual recurring problem, which has actually a different intervention where you kind of look at the underlying belief systems values and the goal is not to solve that problem. The goal in that situation is to actually just get a better understanding of each other and your positions and just you stop seeing your partner as the adversary and you start seeing them as a person who makes sense. But if there's been a specific event, a specific fight that's just sort of situational, but it's left bad blood. Things were said or you didn't feel understood. This intervention I was talking about is one that you would go through a series of steps where first you identify the emotions that you were feeling, then you describe play by play, your movie, your perspective. If your partner were looking through your eyes, this is what they heard, saw, thought, then they saw this, then they heard this. So you're not saying, yeah, then you came in and were yelling and acting crazy. You're saying, so then I saw you come in, I heard you say, and I thought to myself, well, great, now everything's ruined, right? So you're showing them your movie, then they have to summarize the movie for you and then vice versa. And then there's this step where each person validates some part that they can understand. Like based on what you saw, heard, I can actually understand how you felt one of those feelings that you said. And then my favorite part is you rewind sort of the movie from that day back through into childhood and you land on a time, a memory, when you felt a similar set of feelings. And this is like the most beautiful part ever because let's say the feeling was I felt misunderstood. I felt misjudged, uncared about, unloved, like you didn't even like me. And I'll say, when did you feel that way? Land on a time and they're like my whole childhood. My parents were, my mom was always accusing me of doing things I wasn't doing and it would set me up and my dad would come home, he'd hear about it. He would just believe her. And then you have like a partner climbing up on the couch, like gives their partner a hug while they're sharing the story. It's beautiful." } ]

[ { "from": "human", "value": "And it changes the way you interact in future disagreements. She has those moments." }, { "from": "gpt", "value": "Yeah, you can't unlearn. Now you know this about your partner," }, { "from": "human", "value": "you know what they're sensitive to. Again, you kind of see the beauty and the flaws then. Right? It all makes sense, right? Yeah, it all kinda makes sense." }, { "from": "gpt", "value": "Yeah, right. It all makes sense. Yeah, it all kinda makes sense. So you maybe were in this dumpster dive in your head of how your partner sucks and all the things that are wrong with them and it's so hopeless. And then you get this light shining through and you realize, oh my God, of course they would be sensitive to that. And suddenly it's not about all the ways your partner is wrong and proving that they're wrong. It's just, how can I in the future make sure they do not feel this again? I would never want this person I love" }, { "from": "human", "value": "to misunderstand me and feel so unloved. What are you, the early days of that? What do you think about the whole dating, modern dating process? How do you find a partner that you can stay with" }, { "from": "gpt", "value": "for the rest of your life? So we are absolutely doing it wrong. And, but there is a way you can do it. And I am such a fan of the psychologist, Ty Tashiro. I adore him. He's brilliant. He's lovely. He's also very humble. Just a wonderful salt of the earth guy. I'm going to tell you a very true story here. Okay. I was in a bad relationship and I was at a psychology conference with my partner at the time. We were both at this conference and we were sitting in a lecture hall there for Ty Tashiro to do his talk that day on his phenomenal research on relationship satisfaction and dating. And I'm sitting next to him and we'd been, you know it was just always unpleasant on trips. There were always fights. We're sitting there and Ty Tashiro starts talking about his research and how he found that most people are, you know signing this agreement, getting married and doing it based on the love endorphins. And really only about 35% of anybody who's married is actually happy. And he said, so then, you know, and It's a pretty low number. Exactly. But here's what I love about Ty Tashiro is he didn't stop there. He wanted to know what those people who were happy had in common. And then same thing with the people who were unhappy. He found a couple fascinating patterns. So the couples who were happy tended to rate their partners higher in three different traits. And I love talking about this because if you are somebody who can follow instructions you can find this, I mean, very easily. Those three traits tend to be conscientiousness. Okay. And I love the word conscientiousness because it's not just kindness. Kindness is a good way to think of it but you can be kind and kind of be a pushover and that's not attractive. Conscientiousness is smart, attentive. It's somebody who reads into a text message and thinks, wow, she was making herself very vulnerable there." }, { "from": "human", "value": "That's conscientiousness." }, { "from": "gpt", "value": "I like how you're just doing a compliment. It's true. It's a certain intelligence, awareness and attunement. And then on top of that, conscientiousness is motivated. So you can't be on your ass all day and be conscientious because then you can't meet the needs that you anticipate about the person. So conscientious is that guy who drives the car around in the rainstorm so his wife's hair doesn't get met. It's my husband who checks my alarm for me every morning because he knows I'm terrible at time management and he makes sure that I set it a reasonable amount of time before my first meeting and not the 20 minutes I think I need. And then he'll come wake me up with a cup of coffee. That is ultimate conscientiousness. And it is true. I mean, I will tell you as somebody who's with a conscientious partner, your love increases over time as you continue to feel grateful and admiring of that person. The second one, you want somebody who is low in a big five personality trait called neuroticism. You want somebody emotionally stable in a way. Now this doesn't mean you can't have somebody who doesn't get the blues or struggle with mental health issues. Trust me, Ty is with somebody who I'm all over the place. But you want somebody who kind of owns their shit and isn't going to just be emotionally unstable all over. You want somebody who is generally happy and has some life satisfaction. Having a partner who has serious, not mental health issues, but unmitigated emotional distress and instability is really hard on the partner. And it's really hard on other family members, including children if you have children. So it's just a predictor of happiness." }, { "from": "human", "value": "So there's a certain threshold of chaos that if you exceed it, it's going to be destructive" }, { "from": "gpt", "value": "to a long-term relationship. Exceed it is going to a long-term relationship. A perfect description about chaos. Not the mystery chaos you love with your little poet brain. I'm talking more like just somebody who, there's just no peace. There's no peace. There's a problem with everything. Everything becomes more difficult. Going to a party is a chore. You don't know if they're gonna have a meltdown at the party or how many complaints about your friends or everything is a problem. So you want somebody who has just some resiliency, I think is a good term for it." }, { "from": "human", "value": "Some flexibility. But then again, yeah, spice is okay, but not too much." }, { "from": "gpt", "value": "Right, right. Flexibility, resiliency, easy going. Yeah. Okay. The third is really interesting, I think. So he found that having a partner with sort of moderate adventurousness, not high adventurousness actually leads to greater satisfaction. And the reason for that is high adventurousness equals novelty seeking, shiny new things. And so if you're in a monogamous relationship, if that is what's important to you, it's going to be very hard for a partner who is novelty seeking to be faithful. So that will cause a lot of pain. But also novelty seeking people tend to always have new projects, new interesting things. And so their attention is drawn away from the relationship." } ]

[ { "from": "human", "value": "And so you can just feel pretty neglected or unimportant." }, { "from": "gpt", "value": "By a little bit. But you want a little bit of adventurousness. So you want your person to be sort of self-motivated, individuated, have their own interests, not completely dependent on you. But also, I mean, low and adventurousness is not a bad thing. Ultimately, what you're getting with low to moderate adventurousness is that rock, that feeling of stability, that home. And I made some references earlier, like when you're 70 and you turn to your partner, do you want them to be hot? Or, for instance, my dad has dementia right now. And my husband turned to me on the plane, we were all coming back from a trip and where we really saw how severe it's getting. And he just turned to me, he knew how much pain I was in, even though I wasn't showing it. And he said, I want you to know that if it comes to a point where we need to take care of your dad, he needs to live with us, you don't even need to ask. It is, I am 100% on board and will help. And those are the things that matter. That home feeling. And technically, that's a trait that's usually, that's sort of a, my husband caring so much about family and home and taking care of things that matter, those are things that tend to be associated with that low to moderate adventurousness." }, { "from": "human", "value": "Somebody who really cares about simple things and family. And I wonder if those three things are something you can work on, consciousness, you can probably- You can. Proactively observe yourself, and do it more regularly. Right Neurodicism might be the hardest one, probably." }, { "from": "gpt", "value": "I think so. To control." }, { "from": "human", "value": "Well, I mean, I was pretty neurotic in my early 20s. And when you wake up, and maybe if you're self aware about it, maybe you'll be able to control it." }, { "from": "gpt", "value": "Yeah, I think so. Maybe you'll be able to control it. I think self awareness is key. I think that's why I love therapy so much. I think life is about growth and our potential for growth and to make our own lives better, to make the lives of others better, to serve others to heal, all of us are this collective healing" }, { "from": "human", "value": "and I think we're all capable of growth. And the same with adventurousness. I'm somebody that's pretty low on adventure but I keep throwing myself out there" }, { "from": "gpt", "value": "just for the extra adventure so you can grow in that way, the extra adventure so you can grow in that way. Yes, and I am high in adventurousness and I was not really ready to settle down. I was married earlier in my 20s, but I would say that I am much more prepared to be in a committed long-term relationship now in my 40s" }, { "from": "human", "value": "than I was when I was younger. But in that same way for me, I like to connect myself to high adventure people so that it like brings me out. It's like they're a horse and I would get to ride them." }, { "from": "gpt", "value": "Yeah, and that's the thing. So high adventure people are attractive. They're interesting, exciting, but it can be a world of heartbreak because you're only under that spotlight for a few minutes" }, { "from": "human", "value": "and then they're onto the next shiny thing. Yeah, but heartbreak is part of love. But that might be the drug thing that you were talking about. Speaking of adventurousness, what about sex?" }, { "from": "gpt", "value": "What role does sex play in a successful relationship? What role does- Oh, sex is important. Well, okay, so I'm saying it's important, but I wanna qualify that. Everybody has different levels of sex that are satisfying to them. Sex can definitely bond you to your partner orgasms are amazing. They de-stress us. They're healthy. I mean, you can have an orgasm and have a lower level of stress for 48 hours. I think that's pretty incredible. If you have, I mean, just that kind of physical contact with your partner, even a 20 second hug with your partner has similar benefits to an orgasm. You're gonna have a lower stress level. You're gonna feel immediately close to your partner. You're gonna get a rush of oxytocin, which is gonna make you feel happier, more grounded throughout the day. So that's a 20 second hug." }, { "from": "human", "value": "You extrapolate that to sex and things are gonna be great. So it's just physiological. But I wonder, there's probably metrics about how often you have sex," }, { "from": "gpt", "value": "how that correlates to successful relationships and so on. Well, there are, but it really has more to do. It's sort of like, remember I was talking about processing conflict and what matters is, do people feel like it's been resolved? Do they feel like there's been a repair? Not necessarily how they go about doing it. Same with sex. Do, does each partner feel sexually satisfying? So that could be once a month for one couple. It could be five times a week for another couple. It could be never for other couples, truly. I mean, so sex has a ton of benefits, but its absence isn't necessarily detrimental, I guess," }, { "from": "human", "value": "would be the qualifier, depending on who you are. And I know couples, they use sex" }, { "from": "gpt", "value": "as part of the conflict resolution process. It's hugely effective for that. If it works for both parties, all parties," }, { "from": "human", "value": "not just both, all, all parties, not just both, all. That's true. What do you think about infidelity? What's the cause of infidelity?" }, { "from": "gpt", "value": "Why do men and women cheat? It's different for everybody. But I mean, even earlier I was saying with adventurousness, like if monogamy is something you're doing. I've seen in my own practice, I've seen the entire range of couples who are open about having sexual relationships with other people and fine with it. Couples who wanna be fine with it, but find out they're not. Couples who aren't just couples, couples with multiple romantic relationships. I've had couples where affairs are tolerated and not talked about. They're not enjoyed, but they are not the type of betrayal that will destroy the relationship, a sort of a understanding and keep it out of my face." } ]

[ { "from": "human", "value": "And then also we won't talk about it. So in the fair that happened without getting permission first, and as long as you don't talk about it, it's not going to do a damage to the relationship." }, { "from": "gpt", "value": "Right, but we can't even talk about it like that, right? So nobody's going to admit that the fair is happening. There can't be any evidence of it. It's sort of a just look the other way type of a situation. But the partner who is not having the affair, right? They typically know, they certainly know that their partner is capable of that. They just kind of know, but they don't want it in their face. It would become a problem if it was in their face. As long as certain needs are met and everything else is okay at home," }, { "from": "human", "value": "it's just one of those things where don't ask, don't tell. Well, that's an interesting point because I've had a bunch of arguments with people. I tend to hang out with, especially in the tech sector, people who really value like honesty and radical honesty. And I keep arguing with people about this because to me, it's not that simple. Like that's an example right there that honesty can be really destructive. Like honesty is also a really complicated thing to get to the bottom of because what is really honest? Yes. And like, how do I look in this dress? Like there's a million ways to answer that." }, { "from": "gpt", "value": "It can be a yes. Like there's a million ways to answer that. It can be a cesspool in my mind. If I'm in a bad place or my partner and I am, like if Ty have and I haven't been connected lately, my honesty of what I actually think about him would be horrifically damaging and completely unfounded also. And it can change on a dime." }, { "from": "human", "value": "And change on a dime. But that's also not actual honesty to the big picture of how you feel about him. I have interacted with a few folks who talk about their previous sexual partners, for example, the numbers of sexual partners they've had. And they feel like that kind of honesty is actually empowering and enriching to the relationship because all the sexual experiences you've had in the past make you a better sexual partner, a better partner in the present. And to me, from the culture I've come from, that's like anti-romantic. Yep, yep. Like you kind of throw the past kind of away. You don't really talk about it. It's kind of there in this amorphous shape, but it's almost as if you've met together for the first time and this is a beautiful new thing, like your creatures that have woken up from a long slumber. Right, you're starting anew. Starting anew. And you want some mystery there. Right, I think the mystery, and like you have to figure that out about each other. So I'm not exactly sure that honesty is always" }, { "from": "gpt", "value": "a relationship. Yep, right, you're starting anew. Right, it's different for everyone. And then also is honesty harmful or helpful at certain points too? Yeah. So you're talking about sort of like disclosing prior sexual history. I thought you were going to go to, so if you've had an affair," }, { "from": "human", "value": "do you hold, do you keep that under your head?" }, { "from": "gpt", "value": "Oh yeah, that's a really tough question." }, { "from": "human", "value": "Or are you obligated to disclose it?" }, { "from": "gpt", "value": "It's a really, that's a really nice." }, { "from": "human", "value": "It is a very tough question, very tough." }, { "from": "gpt", "value": "Well, what do you think is the right answer? I have my own personal beliefs. I also, then like I have my therapeutic beliefs. I think, frankly, and this is just me as a human being, not Shannon, the psychologist. I believe that if you have fucked up, and again, I'm coming from a framework right now of monogamy, if you are committed to somebody you love and you have fucked up, you don't get to shed your guilt onto them. You need to carry that burden. It's not necessarily, I think it's simplistic and unsophisticated to be like, but then you're being dishonest. I think it's actually selfish to unload it on somebody else and give them the trauma of imagining. What we do know about infidelity is it can create an actual post-traumatic stress-like experience for the betrayed partner, where they are having intrusive thoughts about it. Those are unwanted thoughts and it's uncontrolled. It comes in at multiple times a day. They'll have depressed mood. They'll have nightmares about it. Their entire sense of security, safety, self-esteem gets shattered because of your actions. I think it's kind of, yeah, moralistic and naive to think, well, they deserve to know the truth. If you actually know the harm that that sort of betrayal does, especially if you truly mean to stop it, right? So if it was a one and done or if it happened and you've stopped it and you do not intend to do it again, frankly, I think you live with that burden." }, { "from": "human", "value": "You live with that discomfort. Thank you for saying that. Cause I totally agree, but it's like logically it doesn't quite make sense to give that advice, but psychologically it makes complete sense because you really are destroying another person's mind, their faith and love in relationships, their trust, everything. And then you're imprisoning them to be stuck with you for months or years if you're trying to work through it, through that torture. So you should be carrying that burden and working through it, I think. Why do you say that that's your personal opinion, University of Therapeutic, like what?" }, { "from": "gpt", "value": "Well, I think everybody has different values. So I think that's a value-based decision because to me, the hierarchy is kindness and do no further harm over, in that case, over truth, right? Whereas other people, my husband, for instance, he is like truth above all else. You don't get to decide what I know, or you don't get to decide whether or not I can handle that knowledge. So he would even see my determination that I should carry the burden arrogant. Well, why don't you let your partner decide whether or not they, why do you get to choose? I don't know, I think there's value to both arguments. I absolutely see his point, I absolutely see his point. And this, I think is a very humble sort of option. Like you don't get to choose what's better. You just need to give them the information and they can choose. But I think, I don't know, I think it's kinder to hold. I think it's going to cause your conscience more discomfort to hold it. And I think there's sort of a cleansing we do when we share that information. I think in real life, most people disclose it because they can't stand the secret anymore themselves." } ]

[ { "from": "human", "value": "That to me is a selfish act. See this point. I have unemployment applications and so on. And just with friends would ask people, what do you care more about truth or loyalty? Just to get to see how they think about those different questions. And yeah, I was surprised how much variance there is on that." }, { "from": "gpt", "value": "And also conceptually, I bet, I conceptually, I don't think we actually know where we stand until we're faced" }, { "from": "human", "value": "with a situation like that. Yeah, I think people, a lot of people, especially when they're younger, say, especially if they're kind of intellectual. They'll say truth above all else. Yeah, yeah. It's like. You're exactly right. All right, until. It's a platitude. Until you get to hear a truth that truly breaks you. It truly hurts you. It causes suffering to you. And then you realize, or a truth you get to somebody else will cause them suffering. Right. And then, you get to see that suffering destroy their life. and maybe your relationship and so on." }, { "from": "gpt", "value": "And then you're like, oh, until above all else. All right. You're exactly right. All right. Right, right. But yeah, like, should I sit my dad down right now and be like, dad, your dementia, like you have dementia again today. I'm gonna tell you, dad, you're not making sense. No, it's not going to be discussed. We're gonna make them comfortable. And I mean, yeah, I think truth can be a little bit" }, { "from": "human", "value": "of a platitude sometimes. Some of those complexities are all the things involved in the challenges of what makes a relationship work, right? What do you think about open relationships in general? My worldview is such that I see the beauty and value in monogamous relationships just for me, but I don't, I'm also open to the possibility of what works for other people. Have you done any kind of work with the people" }, { "from": "gpt", "value": "in open relationships? Okay. As clients or research? As clients." }, { "from": "human", "value": "Oh yeah, through research. As clients. Is there some interesting differences" }, { "from": "gpt", "value": "between open relationships and monogamous relationships? You know, I think that may have been actually what was behind my question about the satisfaction with them being on the extremes. My hypothesis essentially was, is it because they, if you are really all in, you've worked out some of the kinks. I think I've seen couples who are trying it out, like for the first time, it tends to get a little haywire. There's some excitement in the beginning. Everybody's really excited about it. I think the philosophy makes sense to a lot of people. The science of it makes sense to a lot of people. And, but we have been raised in a society that is pretty monogamous. So there isn't a lot of scaffolding around it. And there's a lot of inner conflict, I think, for people to go away from the values that they've been taught since they were kids. And so jealousy arises a lot. And also it's very difficult to be, I think, as truthful and direct as you need to be, which you're describing in these polyamorous situations where everybody is laid out on the table. So I think that's something that may be practiced. In my own work with clients, I've just noticed that the partners who are happier in these situations who I've worked with," }, { "from": "human", "value": "they are more experienced at it. They seem to have it down. You testified in the Johnny Depp Amber Heard trial based on your role as a clinical and forensic psychologist. It was watched by, I don't know how many people, maybe tens, maybe hundreds of millions of people." }, { "from": "gpt", "value": "What was that experience like?" }, { "from": "human", "value": "Thank God I didn't know that at the time." }, { "from": "gpt", "value": "Were you scared?" }, { "from": "human", "value": "Oh yeah. Given the size of the platform," }, { "from": "gpt", "value": "how many people are watching? How many people are watching? I'm not, scared typically isn't the word. When I testify, I'm always excited and a little trepidatious before I testify because the stakes are so high for everybody's life in that room. This was different. I, anxiety isn't usually my brand and I just skipped anxiety that morning, went straight to terror and I was mad. I was mad at the legal team. I was funny, like I was having all these strong emotions. I couldn't find my bobby pins. I almost started crying because I couldn't find them. I was pretty unhinged that morning and in a way that was really unfamiliar to me. And it was right when I cried because I couldn't find my bobby pins that I realized I needed to get a grip and that I was a professional and that my hair didn't matter even though it ended up mattering, people noticed that it was crazy." }, { "from": "human", "value": "But I got a grip and I went in and I just did my job. And I just- So the terror in the end helped you focus" }, { "from": "gpt", "value": "and do your job well. I think it does and it's a little scary though because I know what fear does cognitively and there is a sweet spot where you want some stress and then you can be really acutely focused and attuned. But then if you go over this threshold, you get sort of that frontal lobe shutdown where you're not thinking clearly and everybody knows that experience from taking a really stressful test at some point like in high school. And then they're going over the answers with the teacher in class later and they're like, how did I miss that question? I know that, you're just in a different state. That's when you have too much stress. I think this day I actually was bordering on too much stress if not clearly in that threshold. But once you're sitting there for a little bit and you're asked the questions, you can kind of go into a routine" } ]

[ { "from": "human", "value": "of just wanting to talk about your work. So what is the work, the job of a forensic psychologist" }, { "from": "gpt", "value": "in that context? In the DEP herd trial, I was serving as an expert witness based on a psychological evaluation of one of the parties. So forensic psychologists can serve the court or in legal matters in a number of ways. They can act as a confidential consultant for an attorney on a case, or they can even assist with jury selection. They might testify without doing an evaluation if they're just coming in to testify about sort of a subject matter. And then they wouldn't be answering specific questions to either of the parties, but just talking more hypothetically about a field area. In this case, because I was ordered to conduct an evaluation, I evaluated one of the parties and then you provide a report to the court with your findings and then you testify" }, { "from": "human", "value": "as to what your findings were. But from my perspective, just watching you, you seem to have held it together really well. So what do you attribute that to? So you said like it calmed down after you were able to ask the question. So to me, if I were just to put myself in this place, it seems like the internet and the world would be very nitpicky about individual words. You're speaking from a place of scientific rigor. And so you have to be very precise with the awarding. I would feel like so much pressure about each single word I choose. Did you feel that pressure" }, { "from": "gpt", "value": "that you had to be extremely precise with the words? Always. The pressure is so high going in to testifying. I think that's where I feel the most pressure is preparing and literally the moment until I start having to answer. And then I don't even have the luxury of thinking about myself because it is so important that that answer be clarified and understandable to the court that that becomes my focus. And that's the godsend is that I can stop thinking about how scary it all is" }, { "from": "human", "value": "because I need to pay attention to explaining something. So if it's okay, I would love to talk to you about the personality assessment test because I think it's actually super fascinating." }, { "from": "gpt", "value": "But your personality assessment inventory or the MMPI-2? You're probably referring to the MMPI-2" }, { "from": "human", "value": "which is one I talked a lot about. MMPI-2, yeah. So maybe can you explain the MMPI-2 seems fascinating. It has like its output, the results, has some basic skills as code types" }, { "from": "gpt", "value": "and just reading through the different. It's so complex." }, { "from": "human", "value": "It's the thing of beauty. Because the human mind is really complicated. Even, you know, depression, schizophrenia, bipolar disorder like all of these things are really complicated. There's like, many of them we don't understand well. There seems to be a huge amount of variance. And yet you have to be able to stitch together a bunch of characteristics that gives you intuition about the unique aspects of each person. You want to be able to have tests that get you closer to identifying the peculiar flaws or beauties of a particular mind. So this seems to do a good job. Just reading through the different. Like when bread buns. When I broke that in time. I testified. That was the best description. I don't know. I'm being poetic, I apologize. That was a beautiful description You have to being part poetic about the human mind. It's not math. It's psychology, okay. So what is the MMPI-2? Like what are we talking about here? Like, it's a questionnaire." }, { "from": "gpt", "value": "When I broke that in time. I testified. That was the best description no that was a beautiful description. Well it's not math okay. Yes, that's a great start. So it is a questionnaire, 567 yes, no questions. I'm gonna tell you what's most beautiful about this test. So they use an empirical keying method to develop it. What that means is that they didn't have a bunch of psychologists get together and say, let's ask them, let's make sure that we identify people who have somatic complaints or physical complaints by asking them questions about like numbness in their hands, nothing like that. What they did instead was they through, like take a thousand questions out a group of people who they know at a certain mental illness and a group of people who didn't have that mental illness. And then they looked for patterns in what the people with mental illness endorsed as yes and no of those random questions. So it would be, for instance, there's a bronze light fixture right there. One of the questions out of the thousand might be, I like light fixtures that are bronze, true or false. And they looked for correlations in the way people would answer to these completely innocuous, just boring questions. So there was no real way that a test taker" }, { "from": "human", "value": "could foresee the point of answering. And so because they can't foresee, it's very difficult to cheat to get to a conclusion" }, { "from": "gpt", "value": "that you're using that you're talking about. Very difficult, and not only that, but you can imagine using that approach, you can then look for patterns for almost any type of response style for any type of personality trait, any type of mental illness. You just get a comparison group and then a group who's using that specific strategy or has that specific mental illness or has that personality trait and you just look for patterns. And there's a scale output of different kinds and there's code types, and there's code types. Yep, so you've got validity scales and those are just fascinating and often one of the most useful parts of this test in forensic context because they show you how a person is approaching the test, how they're answering questions about themselves. So for instance, you can see if they are tired, you can see if they're kind of responding, responding randomly. You can see if they are in an unsophisticated manner trying to make themselves look perfect, but not very nuanced. You can see if they may be deceiving themselves and truly believe that they are perfect, whereas others don't see it that way. You can see if they're exaggerating, you can see if they're exaggerating because they're truly, it's a cry for help. They are are in extreme distress, but they feel as though they need to really punctuate it to get people to notice. Or you can see if they're exaggerating in a way that is, you know, driven for a specific outcome or gain. It's just fascinating and it's the most well-developed assessment we have for a person's approach to answering questions about themselves." } ]

[ { "from": "human", "value": "So it gives you the context of- How honest they're being." }, { "from": "gpt", "value": "The state of the person is their answer. How honest they're being. Yeah, yeah, they're honesty. How forthcoming they're being" }, { "from": "human", "value": "and how accurate they're being. And then the result of the classification" }, { "from": "gpt", "value": "based on the test are these code types. Right, well, so you have these clinical scales as well. You have 10 clinical scales that look for different kind of primary clinical pathology issues. This test doesn't tell you anything good about yourself. At best, it just tells you that you're not responding in a way that is dishonest and that you are kind, not hugely problematic, but there's, you know, it's not looking for strength. So you have these 10 clinical scales that look for variations above the mean of the population in certain areas, anywhere from, you know, depressive symptoms, manic symptoms, physical complaint, anxiety, nervousness, aggression, social engagement, whole scope of human experience. And then there are much more nuanced scales from those, so little sub scales. And then the real power though of the MMPI-2 is in, as you said, these code types. And these code types are additional patterns that have been detected that really can be more defining of a personality. So you look for peaks. There can be either two extreme peaks or three typically that make a code type. And those peaks are higher scores on these personality traits and specific code types can give you a very nuanced picture of a person's general approach to life and their personal relationships," }, { "from": "human", "value": "just their personality. So you can build on top of those code types an understanding, yeah, how that person's going to deal with different kinds of situations. And then there's, by the way, a lot of code types. There are a lot of code types." }, { "from": "gpt", "value": "I was looking at them. They're pretty interesting. There are a lot of code types. I was looking at them. They're pretty interesting." }, { "from": "human", "value": "They are a lot. I want to take this. Truly fascinating. I want to take this test." }, { "from": "gpt", "value": "I wanted to see which one I would- I have given it to some people in my life." }, { "from": "human", "value": "It's just phenomenal. How hard is it on your side" }, { "from": "gpt", "value": "of the table to give the test. Oh, it's easy. You just proctor, you just make sure that somebody, there's no distraction, that they're well rested, they are sitting there and they can" }, { "from": "human", "value": "just take it in front of you. So I guess the question is because the questions are well designed and that it's hard to mess with them." }, { "from": "gpt", "value": "You just give the." }, { "from": "human", "value": "It's very hard to beat it, you just hand it to them." }, { "from": "gpt", "value": "And it's yes and no." }, { "from": "human", "value": "It's yes and no." }, { "from": "gpt", "value": "Okay. But I should also add to this that this test, as much as I love it, and it is the most researched and widely used personality assessment in the world, it is not in and of itself definitive. So you use it like you already have sort of a hypothesis and you use this for clarification. And it has a ton of value for showing somebody's response or their approach, how forthcoming they're being. But other than that, you really need to consider it as a piece of the puzzle. You had said stitched together earlier and that was just one of those points you made" }, { "from": "human", "value": "that was perfect for describing this. There's probably no one perfect test, right?" }, { "from": "gpt", "value": "For personality?" }, { "from": "human", "value": "No. I wonder, especially with advancements of AI, there could be more and more sophisticated ways of measuring, of collecting data about your behavior. Absolutely, there could be. And being able to measure some kind of more productive kind of, especially not in a forensic context, but more in trying to figure out how to improve your lifestyle, improve your relationships, all that kind of stuff. So the results of the test with Amber Heard, if you can speak to the public stuff, your behavior. Absolutely, there could be. You said that the results of Ms. Heard's evaluation supported two diagnoses, borderline personality disorder and histrionic personality disorder. Can you speak to each one of those? What are the basic characteristics" }, { "from": "gpt", "value": "of borderline personality disorder? Sure, well, so right now, the DSM-5, which is sort of the Bible for mental disorders, it's what we go to our diagnostic manual, it classifies personality disorders according to clusters. And cluster B is one that involves the emotionally erratic, interpersonally erratic emotional disorders. And those include histrionic personality disorder, borderline personality disorder, narcissistic personality disorder and anti-social personality disorder. Eventually there's been some research on this and a lot of support for us, eventually moving into a more spectrum type approach to considering personality disorders, where you'd essentially be looking at dysfunction in different domains of somebody's functioning that has persisted over time. And again, the really important part is it seems to be a stable trait part of their personality that it's in their interpersonal relationships, it's in how they handle their own life, their own functioning, their mood, and it's not just situation-based, it seems to be all areas. I don't love the title, histrionic personality disorder. I think its history, it's pretty controversial and there's some misogyny in it. But that all being said, as a servant to the court and somebody who is there to just provide the science as it exists today, my job is to relay, and in this specific case, I was ordered to provide my diagnostic impressions, a diagnosis, and I don't get to decide which diagnosis, whether I like a certain diagnosis or not, ultimately, if the criteria are met, that diagnosis is given. So as we have it right now with the current personality disorder categories, histrionic personality disorder is probably the most controversial. Some people believe that it is narcissistic personality disorder-lite. So a less obvious, a less malicious version of narcissistic personality disorder. And I think that will probably get sussed out if we do move to a more spectrum-based approach cause then you would be describing sort of a personality disorder and then you would add the traits to it. So you know, with issues in interpersonal functioning and et cetera. So you could be a little bit more specific rather than having to just put somebody in a category." } ]

[ { "from": "human", "value": "So that's where things are moving, you're saying? Of NPD, antisocial personality disorder?" }, { "from": "gpt", "value": "That's where things are moving to more of a spectrum with personality dysfunction. then you list the traits that are there. And I think that'll be more accurate, especially there's so much overlap between these personality disorders right now, especially cluster B, it is not uncommon for people to have two or three personality disorders to meet criteria for two or three at the same time. So speaking about borderline personality disorder and hysteionic personality disorder. Personality Disorder can best be thought of as a disorder of instability and impulsiveness. Emotional instability, instability in a person's self-identity, sense of self, instability in a person's relationships. And then underlying all of this is an intense fear of abandonment. Historyonic Personality Disorder is more of a disorder of emotionality, fate, and attention-seeking. This, you know, histrionic disorder typically is known for the dramatics and people who are observing or interacting with somebody with this disorder may even feel themselves almost kind of wanting to turn away. There's a sense of play acting as the person is speaking or engaging with you. Something just feels a little bit disingenuous and a lot of attention seeking, similar to borderline personality disorder, you might see with histrionic personality disorder, attempts to manipulate. However, the motivation with histrionic personality disorder is that attention. Whereas with borderline personality disorder, the underlying motivation for almost everything is to avoid abandonment. So you'll see frantic attempts to avoid abandonment, frantic attempts to keep people close. And those frantic attempts can be really harmful to the person and to others." }, { "from": "human", "value": "But to the person themselves. So the fear of abandonment can result in the very thing you're" }, { "from": "gpt", "value": "afraid of. Right. And there has been some research also to suggest that borderline personality disorder has different types as well. And I think this is really important because in my own work, I have encountered many people with borderline personality disorder in my own life, right? And there are different types, right? I'm thinking specifically of a girl I really love who I've worked with for years who is so self-aware about this and endearing and in my own life. She owns her shit. I can forgive almost anything if somebody just owns their shit. She might lose her temper. She might lash out. She can be erratic, but she will come back and apologize, own it, and accept full responsibility. And not only that, but identify it and make changes. She doesn't want to be harmful. I adore that about her. I think it's an admirable quality more of us could have. That's very different than when you think about it, there are nine different symptoms and you only need five to meet criteria. So depending on which symptoms you have, you might be far more calculated, conniving, manipulative, or you may just be more of the impulsive, kind of messy, emotionally erratic type. And so there's some new research also coming out that's even suggested that among women, those that score higher in some of these more calculated traits of the disorder may actually be, it may be a certain presentation of female psychopathy." }, { "from": "human", "value": "Are some of these personality disorders, again, probably impossible question to answer, but how much of it is nature? How much of it is nurture? Or how much of it is in the genetics and you just can't do much with? Maybe another question, a different way to ask that is, how much can you help that? How much can you become better?" }, { "from": "gpt", "value": "Mm-hmm. That is a tough question. So there's been a ton of change in the way we've thought about the etiology of these personality disorders specific to borderline personality disorder. I think in general, the view is that most people believe that it was associated with neglect or trauma in childhood. While there is a correlation there, there's a correlation between that and many mental health issues, not just borderline personality disorder. We also, there is evidence to support a genetic basis for this personality disorder and there are people who have borderline personality disorder that report no childhood trauma or difficulty. Sometimes things just happen. So I think it's a mix. I think we need to think of it as biopsychosocial, which is generally the answer to most things when you're talking about how a mental health issue comes to be. I certainly think that in most cases, and here's just me speaking personally, again, I think in my own work, in most cases, what I see is that somebody may have some sort of predisposition, then they go through certain life events and learn patterns of behaving that may serve them well as a child in a dysfunctional situation, but end up being very problematic later on. Or they just have enough hardship that that gene, whatever it was, line dormant, that little borderline personality disorder gene expresses itself. And you'll see that with things like schizophrenia, depression, anxiety disorders, there tend to be certain ages where you'll just" }, { "from": "human", "value": "see that expression happen. All right, for the record, it got cold in here, so we upgraded with a blanket. You look cozy. Just as a question for me, just observing the trial, it was interesting that first of all, it was a really raw and honest exploration of an intimate relationship between two people. It was interesting to watch. I suppose I haven't watched that kind of thing. It made me think about what makes for a good relationship. All the many things we've already talked about in this conversation, it was useful for that. But also there was raw recordings of two humans interaction. What did you think about that, that there's recordings?" }, { "from": "gpt", "value": "It's kind of interesting. The act of recording your partner?" } ]

[ { "from": "human", "value": "Yeah. Not the ethics of that or so on, but the fact that you have this data. It made me wonder, like, if I recorded myself, how would I sound? Well, you do record yourself. No, but here with microphones, but when you're in private, you wonder." }, { "from": "gpt", "value": "Well, you do record yourself." }, { "from": "human", "value": "Like, I had a bit of a fight with a friend last week and I wondered which one of us was the asshole. I would love to hear the recording because we were a little bit, I think we were a little bit rude to each other and I wonder how it went wrong." }, { "from": "gpt", "value": "I love that you asked yourself that question. That's so useful." }, { "from": "human", "value": "We made up the next day and I think both agreed to not ever talk about it. I want the data. Just bury it deep." }, { "from": "gpt", "value": "You know, I record my couple sessions and one of the primary purposes of that is so that after, if they start to get nasty with each other in the session, I can stop it and I can say, what was that? And most of the time that what you're describing is so useful because we don't see ourselves. We have no idea that we just came off as critical. We think we're being completely reasonable and thoughtful. You know, whenever somebody is sort of retelling an argument they got and they said, and then, you know, I was just scaring and just asked, I mean, why is there a reason you didn't, you know, something like that. If they can actually see themselves, they realize, no, their jaw was clenched," }, { "from": "human", "value": "their voice was raised. They actually called a name. Sometimes they're shocked, right? Uh, so just a quick, just a link on it. The, you labeled Amber heard as a three six code type, going back to our discussion, which can mean that quote, she's heavily concerned with image, prone to treating others with cruelty, unable to admit responsibility for wrongdoing, and prone to externalizing blame. And then I also went into the, the MMPI to list three six includes anxiety, tension, rigidity, fear of criticism, suppressed hostility, merging impassive or episodic aggression, suspiciousness, egocentricity, what else? Projection. What can you say about that code that is not captured in the different personality disorders? What are we supposed to do that from a forensic psychology perspective? And what, what, what are we supposed to do that in general? Forget the three six in general, these kinds of code types" }, { "from": "gpt", "value": "in that context, in the context of a trial. If I'm understanding you correctly, it's sort of," }, { "from": "human", "value": "what's the point of these code types? Yeah. Um, thank you for asking the question," }, { "from": "gpt", "value": "but I don't know what I'm doing. I just, I just actually honestly really find MMPI too fast. I love that you do. Yeah. I love that you get it because I just, to me, it's such a, uh, it's almost unbelievable that humans created it. Um, but I think that goes back to that empirical key method of creating something, um, that enabled it to be as robust as it is and something that is very difficult to beat, if not impossible. Um, but the code types really, so it depends on, in any forensic case, what really matters is the legal psycho legal questions. So what is the legal question? And then what is the psychologists responsibility in assisting with whatever question they're being asked? And there's some questions we can't answer, some that we can, um, you don't always need to provide a diagnosis. When you're asked to provide a report, it depends on the jurisdiction. It depends on the statute. Some jurisdictions actually require a diagnosis. In this case, I was asked to provide a diagnosis. So when I'm considering a diagnosis, you're integrating multiple different sources of information. You're you're integrating unexamining self-report. You are adding collateral data usually. I wasn't able to obtain collateral interviews in this case and that was the decision of the court. They said no collateral interviews, but typically that would be something that you would add. You're looking at records ideally from birth up until the date that the alleged injury occurred. And I'm speaking now specifically to a personal injury evaluation or something where somebody is claiming that they were harmed psychologically but you want as many records as possible to show how a person functioned before that event occurred and how they functioned after and you want it to show financial functioning, physical functioning, academic functioning, basically where is there evidence that something in their life changed? Where is there evidence that harm occurred other than from what they're telling you? And in addition to all of those records that you're reviewing, in addition to their self-report, then you're also going to give some of these tests, like the MMPI. So the code types are really that strength of the MMPI-2, it gives you really nuanced information about a person's personality. Now again, you're not going to use MMPI-2 or any other test by itself to diagnose someone or, you know, decide that the person is telling the truth, not telling the truth, it is just another piece of data, and when it's working the way it's supposed to it lines up really nicely with all of the other data you're getting, including what you've observed from the person during your interview with them, the information they're giving you, or inconsistencies with information they're giving you, the consistency or inconsistency of their self-report from" }, { "from": "human", "value": "the records, what the records themselves say, etc., etc. It's helping you clarify and clarify and clarify the picture." }, { "from": "gpt", "value": "You have the person. Yep. Just dialing it down more and more." } ]

[ { "from": "human", "value": "You're just making sure that it is as accurate as possible. Okay. So given how huge this trial was, given how eloquent you were, I know you don't think of it that way, but from a public perspective, you were like the star because of how wild you did. That's insane." }, { "from": "gpt", "value": "I mean, you know, wild you did. That's insane." }, { "from": "human", "value": "I'm pretty sure Camille's the star. Camille is also incredible. I've gotten a chance to interact with her. She's somebody that really inspires me by how good she is at her job, how much she loves her job, and how much the fame, the money, whatever, has not affected the basic core integrity of who she is as a human being. So she's also incredible. Okay, what's the takeaway for you personally from the trial? How has it made you a better person? How has it changed or solidified who you are as a psychologist, as a forensic psychologist," }, { "from": "gpt", "value": "clinical psychologist, and so on? Wow, I mean, a lot happened in my life around that trial, leading up to the trial after the trial. So let's tackle forensic psychologists first. Okay, so in terms of forensic psychology, I am grateful to that trial for really strengthening my abilities. The stakes were so high that I took, I was retained about two years prior to the trial. So I really delved deep into the academic side of forensic psychology and making sure that I was adhering as closely as possible to standard practices, best practice recommendations for this specific type of an examination. It was intellectually awesome and challenging. I felt like my brain was on fire for a full year leading up to the trial. And that can be really, really fun. It was just challenging, but I am really proud of the work I did. I think the stakes were really high. It's serious work. It's important that it's done well and accurately." }, { "from": "human", "value": "And I felt really good about it. So have some of those lessons carried through to your practice now to both research and some of the things you're doing" }, { "from": "gpt", "value": "in terms of helping couples? No, my practice hasn't changed that much. This was more just something that was more, it demanded so much more of my time than my typical forensic work does. And personal injury cases or in cases where there is an allegation of trauma or psychological harm tend to be super labor intensive. This given the magnitude and how long it had been going on in the back and forth required a ton of work before the trial as well. So it pulled me away from the practice. I think it's been nice to go back a bit. Okay, so now personally, I've learned some things. I've learned that I need to slow down a little bit. So this took a lot from you? It took a lot, but it was really the culmination. I feel like there are these hoops we jump through again and again, academic challenges that we continue to meet. And then there's a next one and a next one and a next one. And in the beginning, like when you're getting into college or applying to grad schools, you don't really realize this is going to be a never ending thing, especially if I continue with research or forensic work. I love it because it is so academic. You're writing these 75 page reports and with citations and you have to be accurate. It feels like I'm doing giant board exams again and again and again. It never ends, but that feeling, I think you and I were talking about how it's fun to doubt yourself because it pushes you to do better work. But so if you keep having high stakes, you're going to work all the time, work yourself into the ground, constantly be thinking about, oh, this question, I'm not sure if I fully know the answer and all the research behind that. So I should go there and again, super fun. But I don't just do forensic psychology all day. I also own a clinic. I provide therapy. I've been providing therapy for 15 years. So what happens is you have clients who maybe you've stopped seeing, but when they have a crisis in their lives, they reach out to you again, even if it's seven years later. So you've accumulated hundreds of clients who at any given time are going to reach out when they're in crisis, whether or not you're working on a federal case or in Virginia for this. And that is never going to be something easy to grapple with because I feel that I am letting somebody down. I know I am because these are people I genuinely care about and they care about me and they trust me. And I want to be able to be there for them. I know that it's disappointing if I can't be. And it's also very difficult to separate out the professional therapy relationship from loving someone who you've seen through some of the most difficult parts of their lives. And I can explain that to people all day long, but it doesn't necessarily mean that it's going to be easy for someone to accept when I can't talk to them and they just found out their husband was leading a double life or their sister just died. And I can't even get on a call because I'm getting maybe 15 messages like that a day and have to testify and have to run my practice. So I think that was why I needed to slow down. This case, I was doing all of that. And then like the academic load or the work involved was just tremendous. And some stuff happened. Like my dad, he started having his cognitive decline. I got a medical diagnosis that is stress-induced. I really thought I was getting away with it. I really believed that people who talk like a lot about self-care were kind of full of shit and just didn't know how to push themselves. I still believe in pushing ourselves, but I think I kind of traversed into an area without realizing it where I was no longer pushing myself to challenge myself or see what I was capable of. I was almost pushing myself like as a necessity because I didn't know what else to do anymore. Just an obligation. It wasn't even, I wasn't pushing myself to do. The Depp herd case reminded me of that feeling of pushing myself to do something I wasn't sure I was capable of and overcoming that challenge. That was rewarding. But when you're piling that on with like running a business and all these other things and trying to be perfect at all of them, that just starts to become like a feeling of necessity" } ]

[ { "from": "human", "value": "and it's not healthy. That said, you somehow managed to hold it all together to put forward a masterful performance. And like you said, still take care of all these clients because you're the most important person in their lives for many of them. Is there a secret to that? Is there any hacks? Is there a-" }, { "from": "gpt", "value": "I don't get to sleep a lot, I don't get a lot. No, and honestly, it's a work in progress, right? I don't have an answer for, I wouldn't want my life to be any other way. I wouldn't have had the opportunity to work on this case if I hadn't established my practice and had outreach. And so I can't figure out like which piece you take it out without it all crumbling." }, { "from": "human", "value": "But I would love to have a little more downtime. So it all kind of works together and this passion is the fuel that's behind all of it probably. That's probably the reason you haven't lost your mind" }, { "from": "gpt", "value": "quite yet. Maybe, yeah, maybe. Unless, I mean, it depends who you ask." }, { "from": "human", "value": "Unless I- What about the stress of just being in the public eye?" }, { "from": "gpt", "value": "Has that been difficult for you? That's a lovely question. Thank you for asking it because it is nice to talk to you about this because I feel like you probably understand it a little bit. That was something I was absolutely unprepared for. Like I said, I had no idea how many people were watching when I testified. I had no idea and I got off the stand. I kind of staggered to the back room and truly thought about lying down on the floor because I was so exhausted. And I'd been up studying all my stuff, terrified that I was gonna forget some statistic" }, { "from": "human", "value": "about the MMPI-2. It's gonna be so great. It's great for me. It's gonna be great for people to hear this, that you're human. You're two flaws. That's extremely stressful for many, many hours. I wondered how you can sit there for so many hours and stay so focused and listen so well." }, { "from": "gpt", "value": "It's so difficult. Well, I mean, I could talk about that too. The moment I came to, almost like came back to my body and realized where I was and just wanted it to stop and felt like I was burning alive. I just was thinking, I don't wanna do this anymore. I don't wanna do this anymore. Is this gonna stop? And then another question came and I just had to get back to it. But, so after I testified the first time I went in that back room, I might've laid down on the ground. It's kind of a blur. I mean, I might have. I do remember that Wayne Dennison, one of the senior managing partners at Brown Redneck, who is a phenomenal guy and absolutely brilliant. I will be indebted to him for life because I trusted him. I trusted him and that made all the difference in probably how I testified. But he came in the back and he was looking at his phone and he said, you're on the cover of Time, something on Apple News. And I thought, I mean, I really, I thought he was messing with me." }, { "from": "human", "value": "I thought it was his joke way of saying like I did great. Like I was burning, you know, like Apple. You've worked with veterans. What is PTSD in that context? What's the landscape of psychological suffering" }, { "from": "gpt", "value": "that veteran soldiers go through? Well, if we're talking about combat exposure, you're seeing things you're not meant to see. You're seeing the worst of humanity, people harming other people. It's not natural for others to intend to harm us. It's not natural for us to harm others. And this dehumanization can occur that's so troubling and disturbing that people have a hard time living with it later. Or they just feel this ongoing anger. It depends. It depends on the trauma they're exposed to. It depends on, you know, whether their convoy was ambushed by weapons that were purchased from money that was given to this village from the US government. It depends on whether they did something that they have a hard time reconciling outside of war now that they're back home in civilization. Depends on whether they lost a lot of their comrades and feel that guilt of being a survivor. And again, not everybody develops PTSD. It really, it's a mental disorder. It's serious. We talk so much about trauma and PTSD gets thrown around lightly when actually it's very difficult to meet the full criteria for that diagnosis. And many people experience severe trauma in their lives and only about 14% are likely to actually develop PTSD. It's an exception, not the norm. Traumatic stress is absolutely normal after something traumatic happens. You'll likely have nightmares. You'll likely have anxiety. You'll feel depressed because you're a human being and something abnormal happened. But PTSD is a longer standing condition that is significantly impairing in a person's life. And I think we've lost that in some of the sort of narrative in society. It just, everybody has PTSD. But no, you can have traumatic stress. You can be distressed. You can be affected by trauma and not have that particular diagnosis." }, { "from": "human", "value": "PTSD significantly impairs people's lives. How do veterans, how do soldiers who suffer from PTSD or close to that kind of diagnosis begin to heal?" }, { "from": "gpt", "value": "What's the path for healing? Well, I will hand it to the military because I think in terms of working with their active duty service members, they really invest heavily in mental health. The US Department of Defense was one of the first to bring animal assisted therapy into any type of treatment in the early 1900s with bringing farming into certain hospitals and letting veterans help with the farms and brush the horses, which is so advanced because now we have all this research on animal assisted therapy and how beneficial it is. And just looking in the eyes of a dog can increase your pain threshold and speed healing after a cardiac arrest, help people with dementia to ambulate more freely. It's incredible stuff, simple. And the military was ahead of the game on that. And I don't think that's changed. I did my training at a military hospital in Hawaii, Tripler Army Medical Center. It was phenomenal training. And our psych department, there was so much interesting research going on. We had, and it was so integrated. So you might not imagine that the military would be doing this, but we had an acupuncture department. We had a chiropractic department. We had a yoga section. We were doing yoga sessions there. I mean, anything that has evidence to support its efficacy was being utilized. And I think that's pretty cool about our government. They have a lot of funding, so I'm glad they're using it on that. The real challenge I think comes with the large scale need of the veteran population. And they slipped through the cracks. I know that the DOD had a campaign going where they were doing outreach to anybody who served, for instance, in the Vietnam War. The problem is they were trying to get all of these people assessed for PTSD. And it was great. Like they were getting phone calls, mail. It was sort of saying, hey, we know that you served. Come on in, or let's schedule you an exam with a psychologist and just see if you're owed benefits. The idea of it's great. The problem is that they outsourced to this third-party company. They're paying really low rates for a one-hour meeting with a vet. And you don't need to be specifically trained in assessing PTSD. And so you're getting these variations and opinions that are coming through. And I've had clients who, to me, who I've worked with for years, who have clear combat-related PTSD, according to gold standard measures, according to my knowing them and observing their symptoms and how impaired they are. And it is clearly associated with combat, the content of their intrusive thoughts, their nightmares, et cetera. And they are having a one-hour meeting, sometimes by phone, with one of these psychologists who's been contracted by this third-party organization. That's not even enough for me to get through the first few symptom questions on the CAPS-5 assessment for PTSD. But in that hour, the psychologist is saying definitively, no PTSD. And it's been a travesty for some people, especially for those who need an advocate the most. It tends to happen to my veterans who are maybe a little bit less sophisticated in presenting or advocating for themselves, more humble, less, you know, the guys who deserve it the most, right? They're just getting passed over and it's a maze. I'm not quite sure what the solution is though before. I mean, I've worked for government agencies. It's a massive population. I love that the outreach is even happening and trying to get these guys in for assessment. I think we can criticize any system. I'm glad that system is even happening," } ]

[ { "from": "human", "value": "but it still needs to be better. So I've got a chance to interact with a lot of soldiers that served in Iraq and Afghanistan. And now a lot of soldiers from all different kinds of nations in Ukraine went to the front. There's a bond between soldiers like any other. I don't know if you can speak to why do you think that is on the opposite side of PTSD. There's a deep human connection. There's like a love for each other. What is that?" }, { "from": "gpt", "value": "What is that about war and combat that creates that common? Well, you're seeing, we've talked earlier about that vulnerability, right? So I believe that combat, I believe that most survival situations strip away all ego. And I mean, there are a couple of different layers to this, but I have not served in war. So I wanna be cautious here, but from what I know just about psychology and also from my own experience of survival type experiences, when you're with a group of people and all the ego stripped away, nothing else matters. The focus is on the here and now and a specific mission or your day to day, you can get really close. You're very, very vulnerable. And also in my experience, the guys I work with who have served, there aren't a lot of people who understand what they've been through. Not only some of the unspeakable things they've been through in combat, but some of the things that they feel are unspeakable about returning, especially if they are experiencing trauma, a lot of them, some of the things that service members with PTSD are the most reluctant to disclose is the feeling like they may not know if they love their children anymore or their wife. That they don't even know if they can love anymore. That they feel emotionally numb. That they wanna kill someone. That they have a whole lot of racist beliefs and thoughts. There are a lot of things that can be associated with PTSD that aren't as clear or expected. And these guys don't have many people who understand it or they don't think they would," }, { "from": "human", "value": "but a lot of their fellow service members do. And so I'm going back to Ukraine and boy, nothing makes reveals the human condition. In a more pure form than war, especially the kind of war you get in that part of the world, especially the war in Ukraine, which is a very 20th century kind of war. Brutal, brutal. Well, like I mentioned in a few different ways, you're exceptionally successful by I think the best definition of success. You're doing what you love and you're one of the best people in the world to doing it. And so what advice would you give to young people that look up to you that saw you in the trial, which is your most public facing thing and are just looking, young people that are looking to find what they wanna do with their life career wise" }, { "from": "gpt", "value": "or what advice career wise or? I love that question. What would you tell them? I'm gonna tell them something my dad told me. He said to me, Shannon, just pick anything. Pick anything, if you like it at all studying it, just pick it. He was like, look, don't worry about the job. You don't even know all the jobs that exist. Pick something you like, you will make it your own. And that is exactly what happened. I like psychology. I was reading somehow self-help books. It's not like I had this calling where I, looking back, I can actually create that story because I think now it makes a lot of sense that I do what I do. But I was lost and scared. I started studying psychology. I met a professor who was really inspiring, who wasn't even a psychology professor, but he was public policy. I stayed in touch with that professor. He is a dear friend still to this day. That was 20 years ago. We do research together in Mexico, integrative research with public policy officials and environmental engineers. And I get to be the psychologist on the trip. I never ever dreamed that that sort of stuff could happen. I didn't know about forensic psychology. I also wanna warn anybody who's interested in forensic psychology that it's not like you're solving crimes all day and getting calls by the FBI. You are going to be sitting alone in your home office with your husband, bringing you bowls of cereal and reminding you to go to the bathroom because you haven't gotten up in like 24 hours from the computer and you're gonna have papers all around you and you're just gonna write 75 dense pages with citations of like science. It's brutal. It's academic, but it's fulfilling. My friend Franny posted a meme of one of the girls from Glee or something crying and saying, I'm the happiest I've ever been. And she said, it reminds her when I try to convince her to do forensic site, because I think her mind is perfect for it. You have to be strategic and thorough, but it's a slog, but it's wonderful." }, { "from": "human", "value": "It's wonderful. What would you tell? But you're gonna love it. But it's fulfilling. The image of your husband bringing you cereal while you work on the 75 pages is maybe the most romantic thing I've ever heard. So we started on love. Let me ask one last question about the same topic. What's the role of love in this whole thing, in the human condition," }, { "from": "gpt", "value": "in this whole experiment we've got going on on earth? I think it's all there is, like that jewel song." }, { "from": "human", "value": "How does that go?" }, { "from": "gpt", "value": "God, I don't sing it, don't get it in my head. Please don't sing it, please don't sing it. There have been some profound moments in my life where I feel like I am closest to kind of the truth of life or what it's all about. And usually there's this resonating sense of love and ease and love for myself, love for other people, sort of like it's all okay, we're all okay, we're gonna get through this. I liked what you said about the harm caused by the misinformation or negative things being said about you. Because you're right, it harms that bigger picture." } ]

[ { "from": "human", "value": "I think it holds us back, takes us back from that truth. That there's a love that connects all of us and that if you remember about that love, it's all going to be okay." }, { "from": "gpt", "value": "I really hope it's going to be okay. Me too." }, { "from": "human", "value": "I believe it would be thank you so much for talking today Shannon, you're an incredible person. Thank you for everything you do and for everything you stand for and from everything from your text message to just who you are" }, { "from": "gpt", "value": "and for this amazing conversation today Shannon, you're an incredible person. Thank you for everything you do." }, { "from": "human", "value": "Thank you. Thanks for listening to this conversation with Shannon Currie the support the podcasts. please check out our sponsors in the description. And now let me leave you with some more words from Charles Bukowski. Sometimes you climb out of bed in the morning and you think, I'm not going to make it. But you laugh inside, remembering all the times you felt that way. Thank you for listening. Hope to see you next time." } ]

[ { "from": "human", "value": "Also, if you want to work with our team or always hiring, go to luksfreedman.com slash hiring. And now, on to the full ad reads. As always, no ads in the middle. I try to make this interesting, but if you must skip them, please still check out our sponsors. I enjoy their stuff. Maybe you will too. This show is brought to you by Notion, a note-taking and team collaboration tool, my favorite note-taking and team collaboration tool. And they have a new feature, Notion AI, that I've been using and loving, and this thing is probably the best implementation of a system that uses a language model to generate text, because it integrates across the entirety of your note-taking process, and it adds just a giant number of small and big features that help you out, that save a lot of time, but also make everything more fun and creatively sort of inject ideas into your workflow. So just to list some features that can edit the voice and tone of the text you already wrote, so they can rewrite it in a different tone. They can make the text which I love. They can make it shorter or longer. Also, they can simplify the text," }, { "from": "gpt", "value": "which to me is at the core of the writing process" }, { "from": "human", "value": "to support the learning. Make things as simple as possible, but not simpler, as Einstein said. And to have tools that give you ideas how to do that, not necessarily completely automate everything, but give you really powerful ideas on how to get 90% there. This is just brilliant. Also, if there's technical jargon, they can rewrite the text and explain it more simply. What else? They can obviously summarize the text. If you start writing, they can continue your writing. If you're having trouble starting to write and there's a blank page glaring back at you, they can generate based on a topic, a bunch of text to get you going. I mean, there's so many just amazing features. I love it when great, powerful language models or any idea in AI is then injected into a tool that's actually usable and useful and powerful across a number of use cases to a huge number of people. I mean, this is really, really, really exciting. Notion AI helps you work faster, write better, and think bigger, doing tasks that normally take you hours and just minutes. Try Notion AI for free when you go to notion.com slash Lex. That's all lowercase notion.com slash Lex to try the power of Notion AI today. This show is also brought to you by, indeed, a hiring website. I've used it, I continue to use it to hire folks for the teams I've been on, have led from engineering to creative. Everything requires a rigorous, systematic, artistic, all many adjectives I want to use, process to build up an amazing team because there's nothing more important to the success of an endeavor or the success of life or to just your contentment and happiness and joy and fulfillment and a source of meaning than the team that you take on hard challenges of life with, of work with. So you should use the best tools for the job. Of hiring, and you should take hiring very, very, very seriously. Don't overspend on hiring. Visit indeed.com slash Lex to start hiring now. That's indeed.com slash Lex. Terms and conditions apply. This show is also brought to you by MasterClass. $180 a year, gets you an all-access pass to watch courses from the best people in the world in their respective disciplines. One of the people I just recently talked to is Chris Voss. He is a former FBI hostage negotiator. Brilliant guy. Off the mic, I really enjoy talking to him. There is kindness, camaraderie, thoughtfulness, humor, wit, also a certain sort of cultural density and complexity, hailing from New York or whatever that rich, sexy accent is from. It's just really fun to listen to him, to listen to him discuss what he's really good at. That was true on the podcast, and that is very much true in his MasterClass, where he really systematically breaks down his ideas of what it takes to negotiate with terrorists, negotiate with hostage takers, negotiate with bank robbers, but I think the most important thing is negotiate in everyday life, to negotiate in business, relationships, all of that. It's just a really brilliant, concise, clear, actionable advice that he gives. And that's true for almost every single MasterClass they have, and you get access to all of them. Get unlimited access to every MasterClass and get 15% off an annual membership at masterclass.com. Lex. This is the Lex Friedman podcast. To support it, please check out our sponsors in the description. And now, dear friends, here's Sam Harris. What is more effective at making a net positive impact" }, { "from": "gpt", "value": "on the world, empathy or reason? It depends on what you mean by empathy. There are at least two kinds of empathy. There's the cognitive form, which is, I would argue, even a species of reason. It's just understanding another person's point of view. You understand why they're suffering, or why they're happy, or you have a theory of mind about another human being that is accurate, and so you can navigate in relationship to them more effectively. And then there's another layer entirely, not incompatible with that, but just distinct, which is what people often mean by empathy, which is more a kind of emotional contagion, right? Like you feel depressed, and I begin to feel depressed along with you, because it's contagious, right? We're so close, and I'm so concerned about you, and your problems become my problems, and it bleeds through, right? Now, I think both of those capacities are very important, but the emotional contagion piece, and this is not really my thesis, this is something I have more or less learned from Paul Bloom, the psychologist who wrote a book on this topic titled Against Empathy. The emotional social contagion piece is a bad guide, rather often for ethical behavior and ethical intuitions. Oh boy. And so I'll give you the clear example of this, which is we find stories with a single, identifiable protagonist who we can effortlessly empathize with far more compelling than data, right? So if I tell you, this is the classic case of the little girl who falls down a well, right? This is somebody's daughter, you see the parents distraught on television, you hear her cries from the bottom of the well, the whole country stops. I mean, there was an example of this 20, 25 years ago, I think, where it was just wall to wall on CNN. This is just the perfect use of CNN. It was 72 hours of whatever it was of continuous coverage of just extracting this girl from a well. So we effortlessly pay attention to that. We care about it. We will donate money toward it. I mean, it's just, it marshals 100% of our compassion and altruistic impulse. Whereas if you hear that there's a genocide raging in some country you've never been to and never intended to go to, and the numbers don't make a dent and we find the story boring and will change the channel in the face of a genocide, right? It doesn't matter, and it literally, perversely, it could be 500,000 little girls have fallen down wells in that country and we still don't care, right? So it's, you know, many of us have come to believe that this is a bug rather than a feature of our moral psychology. And so the empathy plays an unhelpful role there. So ultimately, I think when we're making big decisions about what we should do and how to mitigate human suffering and what's worth valuing and how we should protect those values, I think reason is the better tool, but it's not that I would want to dispense" } ]

[ { "from": "human", "value": "with any part of empathy either, either. Well, there's a lot of changes to go on there, but briefly to mention, you've recently talked about effective altruism on your podcast. I think you mentioned some interesting statement. I'm going to horribly misquote you, but that you'd rather live in a world, like it doesn't really make sense, but you'd rather live in a world where you care about maybe your daughter and son more than 100 people that live across the world, something like this. Where the calculus is not always perfect, but somehow it makes sense to live in a world where it's irrational in this way," }, { "from": "gpt", "value": "and yet empathetic in the way you've been discussing. Right, I'm not sure what the right answer is there, or even whether there is one right answer. There could be multiple peaks on this part of the moral landscape, but so the opposition is between an ethic that's articulated by someone like the Dalai Lama, or really any exponent of classic Buddhism, would say that the ultimate enlightened ethic is true dispassion with respect to friends and strangers. So the mind of the Buddha would be truly dispassionate. You would love and care about all people equally, and by that light, it seems some kind of ethical failing, or at least a failure to fully actualize compassion in the limit, or enlightened wisdom in the limit, to care more, or even much more, about your kids than the kids of other people, and to prioritize your energy in that way, right? So you spend all this time trying to figure out how to keep your kids healthy and happy, and you'll attend to their minutest concerns, and however superficial, and again, there's a genocide raging in Sudan, or wherever, and it takes up less than 1% of your bandwidth. I'm not sure it would be a better world if everyone was running the Dalai Lama program there. I think some prioritization of one's nearest and dearest ethically might be optimal, because we'll all be doing that, and we'll all be doing that in a circumstance where we have certain norms and laws and other structures that force us to be dispassionate where that matters, right? So like when I go to, when my daughter gets sick, I have to take her to a hospital. You know, I really want her to get attention, right? And I'm worried about her more than I'm worried about everyone else in the lobby, but the truth is I actually don't want a totally corrupt hospital. I don't want a hospital that treats my daughter better than anyone else in the lobby because she's my daughter, and I've bribed the guy at the door, or whatever, or the guy's a fan of my podcast, or whatever the thing is. You don't want starkly corrupt, unfair situations. And when you sort of get pressed down the hierarchy of Maslow's needs individually and societally, a bunch of those variables change and they change for the worse, understandably. But yeah, when everyone's corrupt and you're in a state of collective emergency, you've got a lifeboat problem, you're scrambling to get into the lifeboat. Yeah, then fairness and norms and the other vestiges of civilization begin to get stripped off. We can't reason from those emergencies to normal life. I mean, in normal life, we want justice, we want fairness, we want, we're all better off for it, even when the spotlight of our concern is focused on the people we know, the people who are friends, the people who are family, people we have good reason to care about. We still by default want a system that protects the interests of strangers too. And we know that generally speaking, and just in game theoretic terms, we're all gonna tend to be better off in a fair system" }, { "from": "human", "value": "than a corrupt one. One of the failure modes of empathy is our susceptibility to anecdotal data. Just a good story will get us to not think clearly. But what about empathy in the context of just discussing ideas with other people? And then there's a large number of people, like in this country, you know, red and blue, half the population believes certain things on immigration or on the response to the pandemic or any kind of controversial issue. Even if the election was fairly executed, having an empathy for their worldview, trying to understand where they're coming from, not just in the explicit statement of their idea, but the entirety of like the roots from which their ideas comes, that kind of empathy while you're discussing ideas, what is in your pursuit of truth, having empathy for the perspective" }, { "from": "gpt", "value": "of a large number of other people, of other versus raw mathematical reason. I think it's important, but I just, it only takes you so far, right? It doesn't get you to truth, right? It's not, truth is not decided by democratic principles and certain people believe things for understandable reasons, but those reasons are nonetheless bad reasons, right? They don't scale, they don't generalize, they're not reasons anyone should adopt for themselves or respect epistemologically. And yet their circumstance is understandable and it's something you can care about, right? And so yeah, like, let me just take, I think there's many examples of this that you might be thinking of, but I mean, one that comes to mind is I've been super critical of Trump, obviously, and I've been super critical of certain people for endorsing him or not criticizing him when he really made it patently obvious who he was. If there had been any doubt initially, there was no doubt when we have a sitting president who's not agreeing to a peaceful transfer of power, right? So I'm critical of all of that and yet the fact that many millions of Americans didn't see what was wrong with Trump or bought into the, didn't see through his con, right? I mean, they bought into the idea that he was a brilliant businessman who might just be able to change things because he's so unconventional and so his heart is in the right place. He's really a man of the people, even though he's gold-plated everything in his life. They bought the myth somehow of largely because they had seen him on television for almost a decade and a half pretending to be this genius businessman who could get things done. It's understandable to me that many very frustrated people who have not had their hopes and dreams actualized who have been the victims of globalism and many other current trends, it's understandable that they would be confused and not see the liability of electing a grossly incompetent, morbidly narcissistic person into the presidency. So I don't, which is to say that I don't blame, there are many, many millions of people who I don't necessarily blame for the Trump phenomenon, but I can nonetheless bemoan the phenomenon as indicative of very bad state of affairs in our society. So there's two levels to it. One is I think you have to call a spade a spade when you're talking about how things actually work and what things are likely to happen or not, but then you can recognize that people have very different life experiences and yeah, I mean, I think empathy and probably the better word for what I would hope to embody there is compassion, right? Like really, to really wish people well and to really wish strangers well effortlessly, wish them well and to realize that there is no opposition between, at bottom, there's no real opposition between selfishness and selflessness because why is selfishness really takes into account other people's happiness? I mean, which, do you wanna live in a society where you have everything but most other people have nothing? Or do you wanna live in a society where you're surrounded by happy, creative, self-actualized people who are having their hopes and dreams realized? I think it's obvious that the second society is much better. however much you can guard your good luck." } ]

[ { "from": "human", "value": "But what about having empathy for certain principles that people believe? For example, the pushback, the other perspective on this, because you said bought the myth of Trump as a great businessman. There could be a lot of people that are supporters of Trump who could say that Sam Harris bought the myth that we have this government of the people by the people that actually represents the people as opposed to a bunch of elites who are running a giant bureaucracy that is corrupt, that is feeding themselves, and they're actually not representing the people. And then here's this chaos agent Trump who speaks off the top of his head. Yeah, he's flawed in all this number of ways. He's a more comedian than he is a presidential type of figure. And he's actually creating the kind of chaos that's going to shake up this bureaucracy, shake up the elites that are so uncomfortable because they don't want the world to know about the game they got running and everybody else. So that's the kind of perspective that they would take and say, yeah, yeah, there's these flaws that Trump has," }, { "from": "gpt", "value": "but this is necessary. Yeah, yeah, I agree with the first part. So I haven't bought the myth that it's, you know, a truly representative democracy in the way that we might idealize. And, you know, on some level, I mean, this is a different conversation, but on some level, I'm not even sure how much I think it should be, right? Like, I'm not sure we want, in the end, everyone's opinion given equal weight about, you know, just what we should do about anything. And I include myself in that. I mean, there are many topics around which I don't deserve to have a strong opinion because I don't know what I'm talking about, right? Or what I would be talking about if I had a strong opinion. So, and I think we'll probably get to that, to some of those topics because I've declined to have certain conversations on my podcast just because I think I'm the wrong person to have that conversation, right? And I think it's important to see those bright lines in one's life and in the moment politically and ethically. So yeah, I think, so leave aside the viability of democracy. I'm under no illusions that all of our institutions are worth preserving precisely as they have been up until the moment this great orange wrecking ball came swinging through our lives. But I just, it was a very bad bet to elect someone who was grossly incompetent and worse than incompetent, genuinely malevolent in his selfishness, right? And this is something we know based on literally decades of him being in the public eye, right? He's not a public servant in any normal sense of that term. And he couldn't possibly give an honest or sane answer to the question you asked me about empathy and reason and like how should we, what should guide us? I genuinely think he is missing some necessary moral and psychological tools, right? And this is, I can feel compassion for him as a human being because I think having those things is incredibly important and genuinely loving other people is incredibly important and knowing what all that's about is that's really the good stuff in life and I think he's missing a lot of that. But I think we don't want to promote people to the highest positions of power in our society who are far outliers in pathological terms, right? We want them to be far outliers in, if in the best case, in wisdom and compassion and some of the things you've, some of the topics you brought up and then we want someone to be deeply informed. We want someone to be unusually curious, unusually alert to how they may be wrong or getting things wrong consequentially. He's none of those things. And insofar as we're gonna get normal mediocrities in that role, which I think is often the best we could expect, let's get normal mediocrities in that role, not once in a generation narcissists and frauds. I mean, it's like, we just take honesty as a single variable, right? I think you want, yes, it's possible that most politicians lie at least some of the time. I don't think that's a good thing. I think people should be generally honest, even to a fault. Yes, there are certain circumstances where lying, I think is necessary. It's kind of on a continuum of self-defense and violence. So it's like if you're gonna, if the Nazis come to your door and ask you if you've got Anne Frank in the attic, I think it's okay to lie to them. But Trump, arguably there's never been a person that anyone could name in human history who's lied with that kind of velocity. I mean, it's just, he was just a blizzard of lies, great and small, to pointless and effective. But it's just, it says something fairly alarming about our society, that a person of that character got promoted. And so, yes, I have compassion and concern for half of the society who didn't see it that way. And that's gonna sound elitist and smug or something for anyone who's on that side listening to me. But it's genuine. I mean, I understand that like, I barely have the, I'm like one of the luckiest people in the world and I barely have the bandwidth to pay attention to half the things I should pay attention to in order to have an opinion about half the things we're gonna talk about, right? So how much less bandwidth than somebody who's working two jobs or a single mom who's raising multiple kids, even a single kid, it's just, it's unimaginable to me that people have the bandwidth to really track this stuff. And so then they jump on social media and they get inundated by misinformation and they see what their favorite influencer just said. And now they're worried about vaccines and it's just, we're living in an environment where the information space has become so corrupted and we've built machines to further corrupt it. We've built a business model for the internet that further corrupts it. So it is just, it's chaos in informational terms and I don't fault people for being confused and impatient and at their wit's end. And yes, Trump was an enormous fuck you to the establishment and that was understandable" } ]

[ { "from": "human", "value": "for many reasons. To me, Sam Harris, the great Sam Harris is somebody I've looked up to for a long time as a beacon of voice of reason. And there's this meme on the internet and I would love you to steel man the case for it and against that Trump broke Sam Harris's brain. That there's something is disproportionately to the actual impact that Trump had on our society. He had an impact on the ability of balanced, calm, rational minds to see the world clearly, to think clearly, you being one of the beacons of that. Is there a degree to which he broke your brain? Otherwise known as Trump derangement syndrome," }, { "from": "gpt", "value": "medical, medical, medical, medical condition. Yeah, I think Trump derangement syndrome is a very clever meme because it just throws the problem back on the person who's criticizing Trump. But in truth, the true Trump derangement syndrome was not to have seen how dangerous and divisive it would be to promote someone like Trump to that position of power and in the final moment, not to see how untenable it was to still support someone who a sitting president who was not committing to a peaceful transfer of power. I mean, if that wasn't a bright line for you, you have been deranged by something because that was one minute to midnight for our democracy as far as I'm concerned. And I think it really was but for the integrity of a few people that we didn't suffer some real constitutional crisis and real emergency after January 6th. I mean, if Mike Pence had caved in and decided to not certify the election, right? Literally, you can count on two hands, the number of people who held things together at that moment. And so it wasn't for want of trying on Trump's part that we didn't succumb to some real truly uncharted catastrophe with our democracy. So the fact that that didn't happen is not a sign that those of us who were worried that it was so close to happening were exaggerating the problem. I mean, it's like, you almost got run over by a car, but you didn't. And so the fact that you're adrenalized and you're thinking, oh boy, that was dangerous. I probably shouldn't wander in the middle of the street with my eyes closed, you weren't wrong to feel that you really had a problem, right? And came very close to something truly terrible. So I think that's where we were and I think we shouldn't do that again, right? So the fact that he's still, he's coming back around as potentially a viable candidate. I'm not spending much time thinking about it, frankly, because I'm waiting for the moment where it requires some thought. I mean, it took up, I don't know how many podcasts I devoted to the topic. It wasn't that, I mean, it wasn't that many in the end against the number of podcasts I devoted other topics, but there are people who look at Trump and just find him funny, entertaining, not especially threatening. It's just good fun to see somebody who's just not taking anything seriously and it's just putting a stick in the wheel of business as usual again and again and again and again. And they don't really see anything much at stake, right? It doesn't really matter if we don't support NATO, doesn't really matter if he says he trusts Putin more than our intelligence services. I mean, it doesn't matter if he's on the one hand saying that he loves the leader of North Korea and on the other threatens to bomb them back to the Stone Age right on Twitter. It all can be taken in the spirit of kind of reality televisions. Like this is the part of the movie that's just fun to watch, right? And I understand that. I can even inhabit that space for a few minutes at a time, but there's a deeper concern that we're in the process of entertaining ourselves to death, right? That we're just not taking things seriously. And this is a problem I've had with several other people we might name who just appear to me to be goofing around at scale and they lack a kind of moral seriousness. I mean, they're touching big problems where lives hang in the balance, but they're just fucking around. And I think there are really important problems that we have to get our head straight around. And we need, it's not to say that institutions don't become corrupt, I think they do. And I'm quite worried that both about the loss of trust in our institutions and the fact that trust has eroded for good reason, right? That they have become less trustworthy. They've become infected by political ideologies that are not truth tracking. I worry about all of that, but I just think we need institutions. We need to rebuild them. We need experts who are real experts. We need to value expertise over amateurish speculation" }, { "from": "human", "value": "and conspiracy thinking and just bullshit and bullshit. What kind of amateur speculation we're doing" }, { "from": "gpt", "value": "on this very podcast? I'm usually alert to the moments where I'm just guessing or where I actually feel like I'm talking from within my wheelhouse. And I try to telegraph that a fair amount with people. So yeah, I mean, but it's not, it's different. Like, I mean, you can invite someone onto your podcast who's an expert about something that you're not an expert about. And then you, in the process of getting more informed yourself, your audience is getting more informed. So you're asking smart questions and you might be pushing back at the margins, but you know that when push comes to shove on that topic, you really don't have a basis to have a strong opinion. And if you were gonna form a strong opinion that was this counter to the expert you have in front of you, it's gonna be by deference to some other expert who you've brought in or who you've heard about or whose work you've read or whatever. But there's a paradox to how we value authority in science that most people don't understand. And I think we should at some point unravel that because it's the basis for a lot of public confusion. And for granted there's a basis for a lot of criticism I've received on these topics, where people think that I'm against free speech or I'm an establishment shill or it's like I just think I'm a credentialist. I just think people with PhDs from Ivy League universities should run everything. It's not true, but there's a ton of confusion. There's a lot to cut through to get to daylight there because people are very confused about how we value authority in the service of rationality generally." } ]

[ { "from": "human", "value": "You've talked about it, but it's just interesting the intensity of feeling you have. You've had this famous phrase about Hunter Biden and children in the basement. Can you just revisit this case? So let me give another perspective on the situation of January 6th and Trump in general. It's possible that January 6th and things of that nature revealed that our democracy is actually pretty fragile. And then Trump is not a malevolent and ultra competent malevolent figure, but is simply a jokester. And he just by creating the chaos revealed that it's all pretty fragile because you're a student of history and there's a lot of people like Vladimir Lenin, Hitler who are exceptionally competent at controlling power, at being executives and taking that power, controlling the generals, controlling all the figures involved and certainly not tweeting, but working in the shadows behind the scenes to gain power. And they did so extremely competently and that is how they were able to gain power. The pushback with Trump, he was doing none of that. He was creating what he's very good at, creating drama, sometimes for humor's sake, sometimes for drama's sake and simply revealed that our democracy is fragile." }, { "from": "gpt", "value": "And so he's not this once in a generation horrible figure. Once in a generation narcissist. No, I don't think he's a truly scary, sinister, you know, Putin-like or Hitler, much less Hitler-like figure, not at all. I mean, he's not ideological. He doesn't care about anything beyond himself. So it's not, no, no, he's much less scary than any really scary, you know, totalitarian, right? I mean, and he's- He's more brave in your world than 1984. This is what, you know, Eric Weinstein never stops badgering me about, but, you know, he's still wrong. Eric, you know, I can, you know, my analogy for Trump was that he's an evil chauncey gardener. I don't know if you remember the book or the film being there with Peter Sellers. But, you know, Peter Sellers is this gardener who really doesn't know anything, but he gets recognized as this wise man and he gets promoted to immense power in Washington because he's speaking in these kind of, in a semblance of wisdom, he's got these very simple aphorisms, or it would seem to be aphorisms. He's just talking, all he cares about is gardening. He's just talking about his garden all the time. But, you know, he'll say something, but yeah, you know, in the spring, you know, the new shoots will bloom and people read into that some kind of genius, you know, insight politically. And so he gets promoted and says, that's the joke of the film. For me, Trump has always been someone like an evil chauncey gardener. He's, it's not to say he's totally, yes, he has a certain kind of genius. He's got a genius for creating a spectacle around himself. Right, he's got a genius for getting the eye of the media always coming back to him. But it's only, it's a kind of, you know, self promotion that only works if you actually are truly shameless and don't care about having a reputation for anything that I or you would want to have a reputation for, right? It's like it's pure, the pure pornography of attention, right, and he just wants more of it. I think the truly depressing and genuinely scary thing was that we have a country that at least half of the country given how broken our society is in many ways, we have a country that didn't see anything wrong with that. Bringing someone who's obviously doesn't know what he should know to be president and who's obviously not a good person, right? Obviously doesn't care about people, can't even pretend to care about people really, right? In a credible way. And so, I mean, if there's a silver lining to this, it's along the lines you just sketched, it shows us how vulnerable our system is to a truly brilliant and sinister figure, right? I mean, like, I think we are, we really dodged a bullet. Yes, someone far more competent and conniving and ideological could have exploited our system in a way that Trump didn't. And that's, yeah, so if we plug those holes eventually, that would be a good thing and he would have done a good thing for our society, right? I mean, one of the things we realized, and I think nobody knew, I mean, I certainly didn't know it and I didn't hear anyone talk about it, is how much our system relies on norms rather than laws. Yeah, civility, almost. Yeah, it's just like it's quite possible that he never did anything illegal, truly illegal. I think he probably did a few illegal things, but like illegal such that he really should be thrown in jail for it, you know? At least that remains to be seen. So all of the chaos, all of the, you know, all of the diminishment of our stature in the world, all of the opportunity costs of spending years focused on nonsense, all of that was just norm violations. All that was just, that was just all a matter of not saying the thing you should say, but that doesn't mean they're insignificant, right? It's not that it's like, it's not illegal for a sitting president to say, no, I'm not gonna commit to a peaceful transfer of power, right, we'll wait and see whether I win. If I win, it was, the election was valid. If I lose, it was fraudulent, right?" }, { "from": "human", "value": "But aren't those humorous perturbations to our system of civility such that we know what the limits are, and now we start to think that" }, { "from": "gpt", "value": "and have these kinds of discussions? That wasn't a humorous perturbation because he did everything he could, granted he wasn't very competent, but he did everything he could to try to steal the election. I mean, the irony is he claimed to have an election stolen from him, all the while doing everything he could to steal it, declaring it fraudulent in advance, trying to get the votes to not be counted as the evening wore on, knowing that they were gonna be disproportionately Democrat votes because of the position he took on mail-in ballots. I mean, all of it was fairly calculated. The whole circus of the clown car that crashed into Four Seasons landscaping, right? And you got Rudy Giuliani with his hair dye, and you got Sidney Powell and all these grossly incompetent people lying as freely as they could breathe about election fraud, right? And all of these things are getting thrown out by largely Republican election officials and Republican judges. It wasn't for want of trying that he didn't maintain his power in this country. He really tried to steal the presidency. He just was not competent, and the people around him weren't competent. So that's a good thing, and it's worth" } ]

[ { "from": "human", "value": "not letting that happen again. But, he wasn't competent," }, { "from": "gpt", "value": "so he didn't do everything he could. Well, no, he did everything he could. He didn't do everything that could have been done" }, { "from": "human", "value": "by someone more competent. Right, but the tools you have as a president, You can do a lot of things. You can declare emergencies, especially dang COVID. You could postpone the election. You can create military conflict, any kind of reason to postpone the election. There's a lot-" }, { "from": "gpt", "value": "Well, he tried to do things and he would have to have done those things through other people and they're people who refuse to do those things. They're people who said they would quit. There would quit publicly, right? I mean, again there are multiple books written about the last hours of this presidency, and the details are shocking in what he tried to do and tried to get others to do, and it's awful, right? I mean, it's just awful that we were that close to something, to a true unraveling of our political process. I mean, it's the only time in our lifetime that anything like this has happened, and it's deeply embarrassing, right, on the world stage. It's just like we looked like a banana republic there for a while, and we're the lone superpower. It's not good, right, and so we shouldn't, there's no, the people who thought, well, we just need to shake things up, and this is a great way to shake things up, and having people storm our capital and smear shit on the walls, that's just more shaking things up, right? It's all just for the lulz. There's a nihilism and cynicism to all of that, which, again, in certain people, it's understandable. Frankly, it's not understandable if you've got a billion dollars, and you have a compound in Menlo Park or wherever, it's like there are people who are cheerleading this stuff, who shouldn't be cheerleading this stuff, and who know that they can get on their Gulf Stream and fly to their compound in New Zealand if everything goes to shit, right? So there's a cynicism to all of that" }, { "from": "human", "value": "that I think we should be deeply critical of. What I'm trying to understand is not, and analyze, is not the behavior of this particular human being, but the effect it had, in part, on the division between people. To me, the degree, the meme, of Sam Harris's brain being broken by Trump represents, you're like the person I would look to," }, { "from": "gpt", "value": "to bridge the division. Well, I don't think there is something profitably to be said to someone who's truly captivated by the personality cult of Trumpism, right? There's nothing that I'm gonna say to, there's no conversation I'm gonna have with Candace Owens, say, about Trump that's gonna converge on something reasonable, right? You don't think so? No, I mean, I haven't tried with Candace, but I've tried with many people who are in that particular orbit. I mean, I've had conversations with people who won't admit" }, { "from": "human", "value": "that there's anything wrong with Trump, anything. So I'd like to push for the empathy versus reason, is when you operate in the space of reason, yes, but I think there's a lot of power in you showing, in you, Sam Harris, showing that you're willing to see the good qualities of Trump publicly showing that. I think that's the way to win over the Candace Owens." }, { "from": "gpt", "value": "But he has so few of them. He has fewer good qualities than virtually anyone I can name, right? So he's funny, I'll grant you that he's funny." }, { "from": "human", "value": "He's a good entertainer. Does others look at just policies and actual impacts?" }, { "from": "gpt", "value": "I've admitted that. No, no, so I've admitted that many of his policies I agree with, many, many of his policies. So probably more often than not, at least on balance, I agreed with his policy that we should take China seriously as an adversary, right? And I think, again, that you have to, there's a lot of fine print to a lot of this because the way he talks about these things and many of his motives that are obvious are things that I don't support, but I'm gonna take immigration. I think it's obvious that we should have control of our borders, right? Like I don't see the argument for not having control of our borders. We should let in who we wanna let in and we should keep out who we wanna keep out and we should have a sane immigration policy. So I didn't necessarily think it was a priority to build the wall, but I didn't, I never criticized the impulse to build the wall because if, you know, tens of thousands, hundreds of thousands of people are coming across that border and we are not in a position to know who's coming, that seems untenable to me. So, and I can recognize that many people in our society are on balance the victims of immigration because, and there is in many cases a zero sum contest between the interests of actual citizens and the interests of immigrants, right? So I think we should have control of our borders. We should have a sane and compassionate immigration policy. We should have, we should let in refugees, right? So I did, you know, Trump on refugees was terrible, but no, like I would say 80% of the policy concerns people celebrated in him are concerns that I either share entirely or certainly sympathize with, right? So like, that's not, that's not the issue. The issue is. A threat to democracy and some fun to do it. The issue is largely what you said it was, it's not so much the person, it's the effect on everything he touches, right? He just, he has this superpower of deranging and destabilizing almost everything he touches and sullying and compromising the integrity of almost anyone who comes into his orbit. I mean, so you looked at these people who served as chief of staff or in various cabinet positions. People had real reputations, you know, for probity and levelheadedness, you know, whether you share their politics or not. I mean, these were real people. These were not, you know, some of them were goofballs, but, you know, many people who just got totally trashed by proximity to him and then trashed by him when they finally parted company with him. Yeah, I mean, it's just people bent over backwards to accommodate his norm violations. And it was bad for them and it was bad for our system. And, but none of that discounts the fact that we have a system that really needs proper house cleaning. Yes, there are bad incentives and entrenched interests. And, you know, I'm not a fan of the concept of the deep state but, because, you know, it's been so propagandized, but yes, there's something like that, you know, that is not flexible enough to respond intelligently to the needs of the moment, right? So there's a lot of rethinking of government and of institutions in general that I think we should do, but we need smart, well-informed, well-intentioned people to do that job. And the well-intentioned part is hugely important, right? This gentleman, just give me someone who is not the most selfish person anyone has ever heard about in their lifetime, right? And what we got with Trump was that like literally the one most selfish person I think anyone could name. I mean, and again, there's so much known about this man. That's the thing. It was like, it predates his presidency. We knew this guy 30 years ago. And this is what, to come back to those inflammatory comments about Hunter Biden's laptop, the reason why I can say with confidence that I don't care what was on his laptop is that there is, and that includes any evidence of corruption on the part of his father, right? Now, there's been precious little of that that's actually emerged. So it's like, there is no, as far as I can tell, there's not a big story associated with that laptop as much as people bang on about a few emails, but even if there were just obvious corruption, right? Like Joe Biden was at this meeting and he took this amount of money from this shady guy for bad reasons, right? Given how visible the lives of these two men have been, right, and given how much we know about Joe Biden and how much we know about Donald Trump and how they have lived in public for almost as long as I've been alive, both of them, the scale of corruption can't possibly balance out between the two of them, right? If you show me that Joe Biden has this secret life where he's driving a Bugatti and he's living like Andrew Tate, right, and he's doing all these things I didn't know about, okay, then I'm gonna start getting a sense that, all right, maybe this guy is way more corrupt than I realized. Maybe there is some deal in Ukraine or with China that is just like this guy is not who he seems, he's not the public servant he's been pretending to be, he's been on the take for decades and decades, and he's just, he's as dirty as can be, he's all mobbed up and it's a nightmare, and he can't be trusted, right? That's possible if you show me that his life is not at all what it seems, but on the assumption that having looked at this guy for literally decades, right, and knowing that every journalist has looked at him for decades, just how many affairs is he having, just how much, you know, how many drugs is he doing, how many houses does he have, what are the obvious conflicts of interest, you know? You hold that against what we know about Trump, right? And I mean, the litany of indiscretions you can put on Trump's side that testify to his personal corruption, to testify to the fact that he has no ethical compass, there's simply no comparison, right? So that's why I don't care about what's on the laptop. Now, if you tell me Trump is no longer running for president in 2024 and we can put Trumpism behind us, and now you're saying, listen, there's a lot of stuff on that laptop that makes Joe Biden look like a total asshole. Okay, I'm all ears, right? I mean, it was a forced, in 2020, it was a forced choice between a sitting president who wouldn't commit to a peaceful transfer of power and a guy who's obviously too old to be president who has a crack-addicted son who lost his laptop. And I just knew that I was gonna take Biden in spite of whatever litany of horrors was gonna come tumbling out of that laptop." } ]

[ { "from": "human", "value": "And that might involve sort of, so the actual quote is Hunter Biden literally could have had the corpses of children in the basement. There's a dark humor to it, right? Which is, I think you speak to, I would not have cared. There's nothing, it's Hunter Biden, it's not Joe Biden. Whatever the scope of Joe Biden's corruption is, it is infinitesimally compared to the corruption we know Trump was involved in. It's like a firefly to the sun is what you're speaking to. But let me make the case that you're really focused on the surface stuff, that it's possible to have corruption that masquerades in the thing we mentioned, which is civility. You can spend hundreds of billions of dollars or trillions towards the war in the Middle East, for example, something that you've changed your mind on in terms of the negative impact it has on the world. And that, the military industrial complex, everybody's very nice, everybody's very civil, that's very upfront. Here's how we're spending the money. Yeah, sometimes somehow disappears in different places, but that's the way war is complicated. And it's, everyone is very polite. There's no Coke and strippers or whatever is on the laptop. It's very nice and polite. In the meanwhile, hundreds of thousands of civilians die. Hate, just an incredible amount of hate is created because people lose their family members, all that kind of stuff. But there's no strippers and Coke on a laptop, so." }, { "from": "gpt", "value": "Yeah, but it's not just superficial. It is, when someone only wants wealth and power and fame, that is their objective function, right? They're like a robot that is calibrated just to those variables, right? And they don't care about the risks we run on any other front. They don't care about environmental risk, pandemic risk, nuclear proliferation risk, none of it, right? They're just tracking fame and money and whatever can personally redound to their self-interest along those lines. And they're not informed about the other risks we're running, really. I mean, in Trump, you had a president who was repeatedly asking his generals, why couldn't we use our nuclear weapons? Why can't we have more of them? Why do I have fewer nuclear weapons than JFK, right? As though that were a sign of anything other than progress, right? And this is the guy who's got the button, right? I mean, somebody's following him around with a bag waiting to take his order to launch, right? That is a, it's just, it's a risk we should never run. One thing Trump has going for him, I think, is that he doesn't drink or do drugs, right? And although there's, you know, people allege that he does speed, but, you know, let's take him at his word. He's not deranging himself with pharmaceuticals, at least." }, { "from": "human", "value": "But apart from Diet Coke, but- There's nothing wrong, just for the record," }, { "from": "gpt", "value": "when you push back on that, there's nothing wrong with Diet Coke. Nothing wrong with Diet Coke, yeah. I mean, it's a very large amount. I occasionally have some myself." }, { "from": "human", "value": "There's no medical, there's no scientific evidence that I observed the negatives of, you know, all those studies about aspartame and all of that is," }, { "from": "gpt", "value": "no, I don't know. I hope you're right. Yeah, I mean, everything you said about the military industrial complex is true, right? And it's been, we've been worrying about that on both sides of the aisle for a very long time. I mean, that's just, you know, that phrase came from Eisenhower. It's, I mean, so much of what ails us is a story of bad incentives, right? And bad incentives are so powerful that they corrupt even good people, right? How much more do they corrupt bad people, right? Like, so it's like, you want to, at minimum you want reasonably good people, at least non-pathological people in the system trying to navigate against the grain of bad incentives and better still, all of us can get together and try to diagnose those incentives and change them, right? And we will really succeed when we have a system of incentives where the good incentives are so strong that even bad people are effortlessly behaving as though they're good people because they're so successfully incentivized to behave that way, right? That's, so it's almost the inversion of our current situation. So yes, and you say I changed my mind about the war. Not quite. I mean, I was never a supporter of the war in Iraq. I was always worried that it was a distraction from the war in Afghanistan. I was a supporter of the war in Afghanistan. And I will admit in hindsight, that looks like, you know, at best a highly ambiguous and painful exercise, you know, more likely a fool's errand, right? I was like, that, you know, it did not turn out well. It wasn't for want of trying. I don't, you know, I have not done a deep dive on all of the failures there. And maybe all of these failures are failures in principle. I mean, maybe it's just, maybe that's not the kind of thing that can be done well by anybody, whatever our intentions. But yeah, the move to Iraq always seemed questionable to me. And when we knew the problem, the immediate problem at that moment, you know, Al-Qaeda was in Afghanistan and, you know, and then bouncing to Pakistan. Anyway, you know, so yes, but my sense of the possibility of nation building, my sense of, you know, in so far as the neocon spirit of, you know, responsibility and idealism, that, you know, America was the kind of nation that should be functioning in this way as the world's cop. And we have to get in there and untangle some of these knots by force rather often, because, you know, if we don't do it over there, we're gonna have to do it over here kind of thing. Yeah, some of that has definitely changed for me in my thinking. I mean, there are obviously cultural reasons why it failed in Afghanistan. And if you can't change the culture, it's, you're not gonna force a change at gunpoint in the culture, it certainly seems that that's not gonna happen. And it took us, you know, over 20 years" } ]

[ { "from": "human", "value": "to apparently to realize that. That's one of the things you realize with the war is there's not going to be a strong signal that things are not working. If you just keep pouring money into the thing," }, { "from": "gpt", "value": "a military effort thing. Well, also there are signs of it working too, you have all the stories of girls now going to school, right, you know, the girls are getting battery acid thrown in their faces by religious maniacs. And then we come in there and we stop that. And now girls are getting educated and there's, I mean, and that's all good. And our intentions are good there. And I mean, we're on the right side of history there, girls should be going to school. You know, Malala Yosafzai should have the Nobel Prize and she shouldn't have been shot in the face by the Taliban, right? We know what the right answers are there. The question is, what do you do when there are enough, in this particular case, religious maniacs, who are willing to die and let their children die in defense of crazy ideas and moral norms that belong in the seventh century? And it's a problem we couldn't solve, and we couldn't solve it even though we spent" }, { "from": "human", "value": "trillions of dollars to solve it. This reminded me of the thing that you and Jack Dorsey jokingly had for a while, the discussion about banning Donald Trump from Twitter. But does any of it bother you now that Twitter files came out that, I mean, this has to do with sort of the Hunter laptop, Hunter Biden laptop story. Does it bother you that there could be a collection of people that make decisions about who to ban and not? And then that could be susceptible to bias" }, { "from": "gpt", "value": "and to ideological influence. Well, I think it always will be, or in the absence of perfect AI, it always will be." }, { "from": "human", "value": "And this becomes relevant with AI as well. Yeah. Because you have some censorship on AI happening. Yeah." }, { "from": "gpt", "value": "And it's an interesting question there as well. I don't think Twitter is what people think it is, right? And I used to think it was more important when I was on it, and now that I'm off of it, I think it's, I mean, first let me say, it's just an unambiguously good thing in my experience to delete your Twitter account, right? It's like, it is just, even the good parts of Twitter that I miss were bad in the aggregate, in the degree to which it was fragmenting my attention, the degree to which my life was getting doled out to me in periods between those moments where I checked Twitter, and had my attention diverted. And I was not a crazy Twitter addict. I mean, I was probably a pretty normal user. I mean, I was not someone who was tweeting multiple times a day or even every day, right? I probably, I think I probably averaged, right? And I had something like one tweet a day, I think I averaged, but in reality, it was like, there'd be like four tweets one day, and then I wouldn't tweet for the better part of a week. But I was looking a lot because it was my news feed. I was just following 200 very smart people, and I would just wanted to see what they were paying attention to. They would recommend articles, and I would read those articles. And then when I would read an article, then I would thought I should signal boost, I would tweet. And so all of that seemed good. And that's all separable from all of the odious bullshit that came back at me in response to this, largely in response to this Hunter Biden thing. But even the good stuff has a downside. And it comes at just this point of your phone is this perpetual stimulus of, which is intrinsically fragmenting of time and attention. And now my phone is much less of a presence in my life. And it's not that I don't check Slack or check email. And I use it to work, but my sense of just what the world is and my sense of my place in the world, the sense of where I exist as a person has changed a lot by deleting my Twitter account. I mean, I had a, and it's just, it's, and the things that I think, I mean, we all know this phenomenon. I mean, we say of someone that person is too online, right? Like, what does it mean to be too online? And where do you draw that boundary? How do you know? What constitutes being too online? Well, in some sense, just being, I think being on social media at all is to be too online. I mean, given what it does to, given the kinds of information it signal boosts and given the impulse it kindles in each of us to reach out to our audience in specific moments and in specific ways, right? It's like, there are lots of moments now where I have an opinion about something, but there's nothing for me to do with that opinion, right? Like there's no Twitter, right? So, like there's lots of things that I would've tweeted in the last months that are not the kind of thing I'm gonna do a podcast about. I'm not gonna roll out 10 minutes on that topic on my podcast. I'm not gonna take the time to really think about it, but had I been on Twitter, I would've reacted to this thing in the news or this thing that somebody did, right? What do you do with that thought?" }, { "from": "human", "value": "No, I just let go of it." }, { "from": "gpt", "value": "Like chocolate ice cream is the most delicious thing ever. It's usually not that sort of thing, but it's just, but then you look at the kinds of problems people create for themselves. You look at the life deranging and reputation destroying things that people do. And I look at the things that have, the analogous things that have happened to me. I mean, the things that have really bent my life around professionally over the past decade. So much of it is Twitter. I mean, honestly, in my case, almost 100% of it was Twitter. The controversies I would get into, the things I would think I would have to respond to and, like I would release a podcast on a certain topic, I would see some blow back on Twitter. It would give me the sense that there was some signal that I really had to respond to. Now that I'm off Twitter, I recognize that most of that was just, it was totally species. It was not something I had to respond to, but yet I would then do a cycle of podcasts responding to that thing that like taking my foot out of my mouth or taking someone else's foot out of my mouth. And it became this self perpetuating cycle, which I mean, if you're having fun, great. I mean, if it's generative of useful information and engagement professionally and psychologically, great. But, and there was some of that on Twitter. I mean, there were people who I've connected with because I just, you know, one of us DM'd the other on Twitter and it was hard to see how that was gonna happen otherwise. But it was largely just a machine for manufacturing unnecessary controversy." } ]