Balaji Srinivasan: How to Fix Government, Twitter, Science, and the FDA | Lex Fridman Podcast #331

Lex Fridman 0:00 Donald Trump was probably the biggest person ever to be removed from social media. Do you understand why that was done? Can you steel man the case for it and against it?

Balaji Srinivasan 0:09 Everybody who is watching this around the world basically saw, let's say, US establishment or Democrat-aligned folks just decapitate, you know, the head of state-

Lex Fridman 0:21 Mm-hmm.

Balaji Srinivasan 0:22 ... from, digitally, right? Like, just boom, gone, okay? And they're like, "Well, if they can do that in public to the US president, who's ostensibly the most powerful man in the world, what does the Mexican president stand against that?" Nothing. Regardless of whether it was justified on the sky, that means they will do it to anybody. Now the seal is broken. Just like the bailouts, as exceptional as they were and at first everybody was shocked about them, then they became a policy instrument. And now there's bailouts happening, th- every single bill is printing another whatever, billion dollars or something like that.

Lex Fridman 0:56 The following is a conversation with Balaji Srinivasan, an angel investor, tech founder, philosopher, and author of The Network State: How to Start a New Country. He was formerly the CTO of Coinbase and general partner at Andreessen Horowitz. This conversation is over seven hours. For some folks, that's too long, for some, too short, for some, just right. There are chapter timestamps, there are clips, so you can jump around or, like I prefer to do with podcasts and audiobooks I enjoy, you can sit down, relax, with a loved human, animal, or consumable substance, or all three if you like, and enjoy the ride from start to finish. Balaji is a fascinating mind who thinks deeply about this world and how we might be able to engineer it in order to maximize the possibility that humanity flourishes on this fun little planet of ours. Also, you may notice that in this conversation, my eye is red. That's from jujitsu, and also, if I may say so, from a life well-lived. This is the Lex Fridman Podcast. To support it, please check out our sponsors in the description and now, dear friends, here's Balaji Srinivasan.

Lex Fridman 2:21 At the core of your belief system is something you call the, uh, prime number maze. I'm curious, I'm curious. We gotta, we gotta start there.

Balaji Srinivasan 2:28 Sure.

Lex Fridman 2:28 If we can start anywhere, it's with mathematics. Let's go.

Balaji Srinivasan 2:31 All right, great. A rat can be trained to turn at every even number or every third number in a maze to, to get some cheese, but evidently, it can't be trained to turn at prime numbers. Two, three, five, seven, and then 11 and so on and so forth. That's just too abstract, and frankly, if most humans were dropped into a prime number maze, they probably wouldn't be able to figure it out either, you know, th- they'd have to start counting and so on, actually pretty difficult to figure out what the, the turning, you know, rule was. Yet, the rule is actually very simple. And so the thing I think about a lot is just how many patterns in life are, we're just like these rats and we're trapped in a prime number maze, and if we had just a little bit more, you know, cogitation, if we had, you know, a little bit more cognitive ability, a little bit more whether it's, uh, you know, brain-machine interface or just better physics, we could just figure out the next step in that prime number maze, if we could just see it, we could see the grid, right? And that's what I think about, like, that, that's a big thing that drives me, is figuring out how do we actually conceive, understand that prime number maze that we're living in.

Lex Fridman 3:35 So understand which patterns are just complex enough that they are beyond the limit of human cognition.

Balaji Srinivasan 3:41 Yes.

Lex Fridman 3:42 And, uh, what do you make of that? Are the limits of human cognition a feature or a bug?

Balaji Srinivasan 3:49 I think mostly a bug. I admire Ramanujan, I admire, uh, you know, Feynman, I m- I admire these great mathematicians and physicists who were just able to see things that others couldn't. And just by writing it down, you know, that's, that's a leap forward. You know, people talk about it's not the idea, it's the execution, but that's for trivial ideas. For great ideas, for Maxwell's equations or Newton's laws or, you know, quantum electrodynamics or some of Ramanujan's identities, that really does bring us forward, especially when you can check them and you don't know how they work, right? You, you have the phenomenological but you don't have the theory underneath it, and then that stimulates the advancement of theory to figure out why is this thing actually working? That's actually, you know, stat mech, you know, arose in part from the kind of phenomenological studies that were basically being done where people were just getting steam engines and so on to work, and then they kind of abstracted out thermodynamics and so on from that, right? So the, the practice led the theory rather than vice versa. To some extent that's happening in neural networks now as, as you're aware, right? And I think that's, um... So just something that's true and that works, even if we don't know yet, that's amazing and that pulls us forward. So I, I do think that the limits are, are more of a bug than a feature.

Lex Fridman 4:59 Is there something that humans will never be able to figure out about our universe? About the theory, about the practice of our universe?

Balaji Srinivasan 5:06 Yeah, people will typically quote Gödel's Incompleteness with, for such question, and, uh, yeah, uh, there are things that are provably unknowable or provably unprovable, um, but I think you can often get an approximate solution. You know, the, the Hilbert, you know, you know Hilbert's problems, like, we will know, we must know? Uh, at least we should know that we can't know. Push to get at least an approximate solution, push to know that we can't know, at least we push back that darkness enough so that we have lit up that corner of the intellectual universe.

Lex Fridman 5:35 Okay, let's actually take a bit of a tangent and explore a bit, in a way that I did not expect we would, but let's talk about the nature of reality briefly. I don't know if you're familiar with the work of Don Hoffman.

Balaji Srinivasan 5:47 No, I don't. I know Roger Penrose has, like, his Road to Reality series for, like, basic physics getting up to everything we know, but go ahead, tell me-

Lex Fridman 5:54 Oh, it's even wilder.

Balaji Srinivasan 5:55 Yeah.

Lex Fridman 5:55 In modern physics, we start to question of what is fundamental and what is emergent in this beautiful universe of ours, and there's a bunch of folks who think that spacetime as we know it, the four-dimensional space, is emergent, it's not fundamental to the, the physics of the universe, and the same-... many argue, I think Sean Carroll is one of them, is that time itself, the way we experience it, is also emergent. It's not fundamental-

Balaji Srinivasan 6:21 Mm-hmm.

Lex Fridman 6:21 ... to the way our universe works. Anyway-

Balaji Srinivasan 6:23 Hmm.

Lex Fridman 6:23 ... those are, uh, the technical term, I apologize for swearing, those are the mind-fucks of, uh, modern physics. But if we stroll along that road further, we get somebody like Donald Hoffman, who makes the evolutionary case that the reality we perceive with our eyes is not only an abstraction of objective reality, but is actually completely detached.

Balaji Srinivasan 6:47 Hmm.

Lex Fridman 6:48 Like we're in a video game, essentially, that's, uh, consistent between each, uh, consistent for all humans, but it doesn't, it's not at all connected to physical reality.

Balaji Srinivasan 6:58 Hmm.

Lex Fridman 6:58 So it's, it's an illusion.

Balaji Srinivasan 6:59 It's like a version of the simulation hypothesis, is that his...

Lex Fridman 7:02 In a very distant way, but, uh, the simulation says that there's a sort of computational nature to reality, and then there's a kind of, a programmer that creates this reality and so on. No, he s- he says that we humans have a brain that is able to perceive the environment, and, uh, evolution has produced from primitive life to complex life on Earth, produced the kind of brain that doesn't at all need to sense the reality, uh, directly. So like, this table, according to Donald Hoffman, is not there (laughs) .

Balaji Srinivasan 7:34 Well, so that-

Lex Fridman 7:36 Like, not, not just as an abstraction, like we don't sense the molecules that make up the table, but all of this is fake.

Balaji Srinivasan 7:43 Interesting. So, uh, you know, I, I tend to be more of a hard science person, right? And so, um, you know, so just on that, people talk about qualia, you know, like, uh, is your perception of green the different fr- you know, different from my perception of green? And, you know, my counterargument on that is, well, we know something about, you know, spectrum of light, and we can build artificial eyes. And if we can build artificial eyes, which we can, you know, they're, like, they're not amazing, but you can actually, you can do that. You can build artificial ears and so on. Obviously we can build recording devices and, you know, for cameras and things like that. Well, operationally, the whole concept of your perception of green, you see green as purple, I see green as green, or what I call green, doesn't seem to add up because it does seem like we can do engineering around it, right? So the Hoffman thing, I get why people more broadly will talk about a simulation hypothesis because, you know, it's like, uh, Feynman and many others have talked about how, uh, math is surprisingly useful to describe the world.

Lex Fridman 8:42 Yeah.

Balaji Srinivasan 8:42 You know, like very simple equations give rise to these complex phenomena. Wolfram is also on this, um, from a, from a different angle with the cellular automata stuff but, um...

Lex Fridman 8:52 It's almost suspicious how well it works.

Balaji Srinivasan 8:54 Yeah, but on the other hand, it's like, uh, you know, it, it is... yet we're still also in a prime number maze, you know? There's things we just don't understand. And-

Lex Fridman 9:04 Right.

Balaji Srinivasan 9:04 ... um, you know, so-

Lex Fridman 9:05 Oh, so the, within the constraints of the non-prime numbers, we find math to be extremely effective, surprisingly effective.

Balaji Srinivasan 9:13 Yeah, exactly. So maybe, maybe the math we have gets us through the equivalent of the even turns and the odd turns, but there's math we don't yet have that is more complex or more complex rules for other .

Lex Fridman 9:22 Despite all our rage, we're all just-

Balaji Srinivasan 9:24 ... e- and, and I think-

Lex Fridman 9:24 ... rats in a cage.

Balaji Srinivasan 9:25 I know that gets, like, very abstract, but, you know, there are unsolved problems in, in physics, uh, you know, s- like the condensed matter space, there's a lot of interesting stuff happening. My recollection, I may be, you know, out of date on this, like, things like sonoluminescence, we don't know exactly how they work. And sometimes those things that are like at the edges of physics, you know, in the late 1800s, I think Rutherford, uh, s- somebody fam- I think it was Rutherford said, you know, basically all of physics is being discovered, et cetera, and that was obviously before quantum mechanics. You know, that, that sort of edge case people were looking at the Balmer and the Paschen series and seeing, you know, this weird thing, you know, with the, with the hydrogen spectrum and it was quantized and, you know, that led to, uh, like the, the sort of phenomenological set of observations that led to quantum mechanics and, and, and everything. And, you know, sometimes I think the UAP stuff might be like that, right? People immediately go to aliens for UAP, like the unidentified aerial phenomena, right? And people have been... Uh, there's surprising amount of stuff out there on this. The UK has declassified a bunch of material. You know, Harry Reid, who's a senator, has talked about this. It's not an obviously, it's not an obviously political thing. It, it... which is good. It's something that... Is, is there something happening there, right?

Lex Fridman 10:34 Mm-hmm.

Balaji Srinivasan 10:34 And people had thought for a long time that the UAP thing was a, like, American, um, kind of counter-propaganda to cover up their new spy planes that were spying over the Soviet Union, to make anybody who talked about them seem, you know, crazy and, and, and hysterical or whatever. But if the UAP thing is real, it could be atmospheric p- phenomena, like, you know, like the aurora borealis or the northern lights, but some things we don't understand. It could be something like the, uh, the Balmer and Paschen series, uh, you know, which were the observations of, like, emission spectra before quantum mechanics. So that's like another option as opposed to doesn't exist or little green men. It could be physics we don't understand yet. That's one possibility.

Lex Fridman 11:15 Do you think there's alien civilizations out there?

Balaji Srinivasan 11:18 So there's a lot of folks who have kind of written and talked about... There's this, you know, the Drake equation-

Lex Fridman 11:22 Yeah.

Balaji Srinivasan 11:22 ... which is like, you know, the multiplying all the probabilities together. There's p- perhaps more sophisticated takes like the, uh, the dark forest, you know, which says that if the universe is like a dark forest, we're the dumb ones that aren't hiding our presence. Um, there's one calculation I saw, I haven't reproduced it myself, but basically says that the, uh, the assumption that other civilizations have seen ours is wrong because when you have like a spherical radius for like the, you know, electromagnetic radiation that's leaving our planet, as that sphere gets larger and larger, it gets, like, smaller and smaller amounts of energy. So, you know, you get farther out, you're, you're not getting enough, um, you know, uh, you know, photons or, or what have you to, to actually, uh, detect it. Um, I, you know, I don't know, I actually haven't looked into the math behind it, but I remember, remember seeing that argument. So actually, it is possible that it's so diffuse when you go past a certain...... you know, a number of light years out that people, you know, th- that an alien civilization wouldn't be able to detect it, right? That's, that's another argument.

Lex Fridman 12:18 That's more basically about signals-

Balaji Srinivasan 12:20 Signal strength.

Lex Fridman 12:20 ... from them, from us-

Balaji Srinivasan 12:22 Yeah.

Lex Fridman 12:22 ... to be able to... Signals colliding en- enough to, uh, find the signal from the noise.

Balaji Srinivasan 12:28 Right. Exactly.

Lex Fridman 12:28 Intelligent signal-

Balaji Srinivasan 12:30 Yeah. Hanson-

Lex Fridman 12:30 ... from the noise.

Balaji Srinivasan 12:31 Hanson has a article called Grabby Aliens.

Lex Fridman 12:33 Yeah.

Balaji Srinivasan 12:34 Um, have you seen his thing?

Lex Fridman 12:35 Yes, yes, yes, yes.

Balaji Srinivasan 12:35 Right? And so there's-

Lex Fridman 12:36 He's been on this podcast.

Balaji Srinivasan 12:38 Oh, great.

Lex Fridman 12:38 He's brilliant.

Balaji Srinivasan 12:39 I like him. He, he pushes, you know, boundaries in interesting ways.

Lex Fridman 12:42 In every ways. In all of the ways, yeah.

Balaji Srinivasan 12:44 In all the ways. That's right.

Lex Fridman 12:45 Yeah.

Balaji Srinivasan 12:45 I, I, I like him overall. He's, he's, you know, he, he, he's an asset to humanity. He-

Lex Fridman 12:49 Grabby aliens? So, he h- he has this interesting idea that, uh, th- civilizations, uh, quickly learn how to travel close to the speed of light.

Balaji Srinivasan 12:59 Right.

Lex Fridman 12:59 So we're not gonna see them until they're here.

Balaji Srinivasan 13:03 Yeah. That's possible. I mean, th- one of the things is... So here's for example a mystery that we haven't yet done, right? Which, or we haven't really figured out yet, which is, um, abiogenesis in the lab, right? We've d- done lots of things where we- you've got... You can show macromolecules binding to each other. You can show, you know, evidence for the so-called RNA world. Abiogeneses to go from, you know, like non-life to life, right? In the lab. You can show microevolution obviously with bacteria. You can do artificial selection on them. Lots of other aspects of, um, you know, fundamental, you know, biochemistry origins of life stuff have been established. There's a lot of plausibility arguments about the primitive environment and nitrogens and carbons snapping together to get, you know, the a- you know... The RNA world is the, uh, the, uh, the initial hypothesis. But to my knowledge at least, we haven't actually seen abiogenesis demonstrated. Now one argument is you need just like this massive world with, uh, you know, so many different reps before that actually happens, and, um, one possibility is if we could do atomic level, you know, simulations of molecules bouncing against each other, it's possible that in some simulation we could find a path, a reproducible path to abiogenesis and then just, you know, replicate that in the lab, right? Um, I, I don't know, okay? Uh, but that seems to me to be like a mystery that we still don't fully understand, like an example of the prime number maze, right?

Lex Fridman 14:27 One of the most fascinating mysteries.

Balaji Srinivasan 14:29 One of the most important, yep.

Lex Fridman 14:30 Yeah.

Balaji Srinivasan 14:30 Yeah. And, and again, there may be some biochemist who's like, "Oh, Balaji, you didn't know about X, Y, and Z that happened in the abiogenesis field." I, I freely confess I'm not, like, uh, you know, au courant on it. Uh, last thing I remember looking at it is, uh-

Lex Fridman 14:43 What does au courant mean?

Balaji Srinivasan 14:44 Like up to the moment, you know?

Lex Fridman 14:45 Oh, nice. That's a nice word.

Balaji Srinivasan 14:47 That's a, a-

Lex Fridman 14:47 Au courant.

Balaji Srinivasan 14:48 Yeah, I'm probably mispronouncing it, but, um-

Lex Fridman 14:49 Yeah. We'll edit it in post to-

Balaji Srinivasan 14:52 Sure, sure, sure.

Lex Fridman 14:52 ... pronounce it correctly with AI.

Balaji Srinivasan 14:54 Sure. Yeah, yeah, yeah (laughs) . It'll be-

Lex Fridman 14:55 We'll, we'll (laughs) , we'll copy your voice and it will pronounce it perfectly correctly-

Balaji Srinivasan 14:59 Yeah.

Lex Fridman 14:59 ... in post.

Balaji Srinivasan 15:00 One thing that I do think was interesting is, uh, Craig Venter a while back tried to make a minimum viable cell, um, where he just tried to delete all of the genes that were, that were not considered essential, and so it's like a new life form. And this was like almost 20 years ago and so on, and that thing was a- was, was viable in the lab, right? And so it's possible that you could... You kind of reverse engineer... So y- you're coming at the problem from different directions, like RNA molecules can do quite a lot. You've got some r- you know, reasonable assumptions as to how that could come together. Uh, you've got like sort of stripped down minimum viable life forms and, and so on and so forth. There's l- it's not there isn't stuff here. You can see microevolution, you can see at the sequence level, you know, if you do molecular phylogenetics you can actually track back the bases. There's actually... So it's not like there's no evidence here. There's a lot of tools to work with. But this, in my view, is a fascinating area, and actually also relevant to AI because another form of abiogenesis would be if we were able to give rise to a different branch of life form, the silicon-based as opposed to carbon-based, you know, to, to stretch your point. Um, you give rise to something that actually does meet the definition of life for some definition of life, right?

Lex Fridman 16:05 What, what, what do you think that definition is for an artificial life form? 'Cause you mentioned consciousness.

Balaji Srinivasan 16:11 Yeah.

Lex Fridman 16:11 When will it give us pause that we created something that feels by some definition or by some spiritual, poetic, romantic, philosophical, mathematical definition that it is alive-

Balaji Srinivasan 16:28 Right.

Lex Fridman 16:28 ... and we wouldn't want to kill it?

Balaji Srinivasan 16:30 So, couple of remarks on that. One is, um, Francis Crick of, of Watson and Crick, uh, before he died, I think his last paper was published on something called the claustrum, okay? And the thing is that, you know, sometimes in biology or in any, you know, domain, people are sort of discouraged from going after the big, the big questions, right? But he proposed the claustrum is actually the organ that is the seat of consciousness. It's like this sheath that, like, covers the brain, and, uh, for mice, if you... And, and a- again, I may be r- recollecting this wrong, so but you can look, but, uh, my recollection is, um, in mice, if you disrupt this, the mouse is like very disoriented, right? It's like... It, it's the kind of thing which, you know, Watson and Crick were all about structure implies function, right? Found the structure of DNA this amazing thing, and, you know, they remarked in this very under- understated way at the end of the paper that, well, obviously this, uh, gives a basis for how the genetic material might be replicated and error corrected because, you know, helix unwinds and you cut and paste it, right? So he was a big structure function person, and that applies not just at the protein level, not just at the level of DNA, but potentially also at the level of organs. Like the claustrum is kind of this system integrated level, right? It's like the, the last layer in the neural network or something, you know? Um, and, uh, and so that's, that's the kind of thing that I think is worth studying. Um, so consciousness is another kind of big qu- abiogenesis, a big question. The prime number maze consciousness is a big question. Um, and, uh, you know, then definition of life, right? Uh, there's folks... Gosh, there's... I think... I... So this one is something I'd have to Google around, but there's a guy I think at Santa Fe Institute or something who had some definition of life and like some thermodynamic definition. Um, but you're right that it's gonna be a multi-feature definition. We might have a Turing test like definition frankly, which is just if enough humans agree it's alive, it's alive.... right? And that might frankly be the operational definition, uh, 'cause, you know, viruses are like this boundary case, you know, are they, are they alive or not? Most people don't think they're alive, but they're, they're, they're on, they're kind of, they're more alive than a rock in a sense.

Lex Fridman 18:35 Well, I think, uh, in a world that we'll talk about today quite a bit, which is the digital world, I think the most fascinating philosophically and technically definition of life is life in the digital world.

Balaji Srinivasan 18:49 Mm-hmm.

Lex Fridman 18:50 So chatbots, essentially creatures, whether they're replicas of humans or totally independent creations, perhaps in an automatic way, I think there's going to be chatbots that we would ethically be troubled by if we wanted to kill them. They would have the capacity to suffer. They would be very unhappy with you trying to sh- turn them off.

Balaji Srinivasan 19:17 Hmm.

Lex Fridman 19:18 And then there would be large groups of activists that will protest and they'll go to the Supreme Court of whatever the Supreme Court looks like in t- in, uh, 10, 20, 30, 40 years, and they will, uh, demand that these chatbots would have the same rights as us humans. Do you think that's possible?

Balaji Srinivasan 19:37 I saw that Google engineer who was basically saying this had already happened, and I, uh, I was surprised by it, uh, because it just, I, when I looked at the chat logs of it, it didn't seem particularly interesting. On the other hand, I can definitely see... I mean, GPT-3, for people who, you know, haven't paid attention, shows that serious step-ups are possible, and obviously, you know, um, y- you've talked about AI on your podcast a ton. Um, is it possible that GPT-9 or something is, is kind of like that? Or GPT-15 or, or GPT-4 maybe? But-

Lex Fridman 20:14 Yeah, for people just listening, there's a deep skepticism in your face.

Balaji Srinivasan 20:19 Yeah, you know, the reason being because, um, i- you know, it's possible. It's possible that you have, like, a partition of society on literally this basis, you know?

Lex Fridman 20:30 Mm-hmm.

Balaji Srinivasan 20:30 Um, that's one model, where there's some people... Just like there's vegetarians and non-vegetarians, right? There may be, um, machines have life and machines are machines, you know? Like, or something like that, right? Uh, you know, you could, you could definitely imagine some kind of partition like that in the future, where your fundamental political, social system, that's a foundational assumption. And, you know, is AI... Does it, you know, deserve the same rights as, like, a human or... For example, a corporation is an intermediate. Uh, do you see that thing which is how human i- are different corporations? Have you seen that infographic?

Lex Fridman 21:10 (laughs)

Balaji Srinivasan 21:11 It's actually funny. So it's like-

Lex Fridman 21:12 Yeah, there's a spectrum.

Balaji Srinivasan 21:12 There's a spectrum. So for example, Disney is considered about as human as, like, a dog. But, like, Exxon... I, I may be , but they had, like, a level with, like, human at one end and, like, rock at the other.

Lex Fridman 21:22 Does it have to do with corporate structure? What, what, what's the-

Balaji Srinivasan 21:24 I think it's about people's empathy for that corporation, their brand-

Lex Fridman 21:27 Oh.

Balaji Srinivasan 21:27 ... identity. But it's interesting to see that, first of all, people sort of do think of corporations as being more or l-

Lex Fridman 21:35 Yeah.

Balaji Srinivasan 21:35 Like, the branding is really what they're responding to, right?

Lex Fridman 21:37 Well, that's what... I mean, they're also responding... You know, I have a brand of human that I'm trying to sell.

Balaji Srinivasan 21:43 Mm-hmm.

Lex Fridman 21:43 And it seems to be effective for the most part.

Balaji Srinivasan 21:46 (laughs) Sure.

Lex Fridman 21:46 Although it's, it has become, like, a running joke that I might be a robot.

Balaji Srinivasan 21:50 Right.

Lex Fridman 21:51 Which means there's, the brand is cracking. (laughs)

Balaji Srinivasan 21:54 Could be.

Lex Fridman 21:55 It's seeping through. But I mean, in that sense, I just, I think, uh, I don't see a reason why chatbots can't manufacture the brand of being human, of being sentient.

Balaji Srinivasan 22:09 I mean, that is the Turing test, but it's like the multiplayer Turing test. Now that actually a fair number of chatbots have passed the Turing test, I'd say there's at least two steps up, right? One is, um, a multiplayer Turing test where you have chatbots talking to each other and then you ask, "Can you determine the difference between end chatbots talking to each other and clicking buttons and stuff in apps and, and humans doing that?" And I think we're very far off. I shouldn't say very far off. At least I don't know how far off we are in terms of time, but we're still far off in terms of a group of end chatbots looking like their digital output is like the group of end humans. Like a girl from the Turing test to the multiplayer Turing test. That's one definition. Another definition is, you know, to be able to kind of swap in and you're not just convincing one human that this is a human for a small, you know, session. You're convincing all humans that this is a human for end sessions. Remote work actually makes this possible, right? That's another definition of a multiplayer Turing test, where basically you have a chatbot that's fully automated, that is earning money for you as an intelligent agent on a computer that's able to go and get remote work jobs and so on. I would consider that next level, right?

Lex Fridman 23:23 Mm-hmm.

Balaji Srinivasan 23:24 If you could have something that was like that, that was competent enough to... I mean, because everything on a computer can be automated, right? Literally, you could be totally hands-free. Just like autonomous driving, you could have autonomous earning. As a challenge problem, if you were Microsoft or Apple and you had legitimate access to the operating system, just like Apple says, "Can you send me details of this event?" A decentralized thing could, in theory, log, you know, the actions of 10,000 or 100,000 or a million people, and with cryptocurrency, you could even monitor a wallet that was on that computer and you could see, you know, what long run series of actions were increasing or decreasing this digital balance. You see what I'm saying?

Lex Fridman 24:05 Mm-hmm.

Balaji Srinivasan 24:06 Right? So you start to get...... at least conceptually. It'd be invasive and, and, you know, there'd be a privacy issue and so on. Conceptually, you could imagine an agent that could learn what actions humans were doing that resulted in the increase of their local cryptocurrency balance, okay?

Lex Fridman 24:21 Mm-hmm.

Balaji Srinivasan 24:22 The- there may be better ways to formulate it, but that I would consider a challenge problem is to go from the Turing test to a genuine intelligent agent that can actually go and make money for you. If you can do that, that's a big deal.

Lex Fridman 24:31 Oh.

Balaji Srinivasan 24:31 People obviously have trading bots and stuff, but that would be, you know, the next level. It's typing out emails, it's creating documents, it's actually-

Lex Fridman 24:36 So it'll mimic human behavior in its entirety?

Balaji Srinivasan 24:39 Yeah. That's right. And it can, it'll schedule Zooms, it'll send emails, it'll essentially... 'Cause if you think about it, a human is hitting the keys and clicking the mouse. But just like a self-driving car, the wheel rotates by itself, right? Those keys are effectively just... It's like a, like the Automator app in, in Apple, right? Um, everything's just moving on the screen. You're seeing it there, and it's just an AI.

Lex Fridman 25:00 It's kind of hilarious that the "I'm not a robot" click thing-

Balaji Srinivasan 25:04 Mm-hmm.

Lex Fridman 25:04 ... actually works. 'Cause I, I, I actually don't know-

Balaji Srinivasan 25:08 How, how that's happening?

Lex Fridman 25:08 ... how it works, but I think it has to do with the movement of the mouse-

Balaji Srinivasan 25:13 Yes.

Lex Fridman 25:13 ... the timing, and they know that it's very difficult for... currently for a bot to mimic human behavior in the way they would click that little checkbox.

Balaji Srinivasan 25:22 Yeah, exactly. I think it's something... I mean, uh, m- again, my recollection on that is it's like a pile of highly obfuscated JavaScript with all kinds... It looks like a very simple box, but it's doing a lot of stuff and it's collecting all kinds of instrumentation. And yeah, exactly like a, like a robot is just a little too deterministic, or if it's got noise, it's like Gaussian noise, and the way humans do it is just not something that you'd easily be able to do without collecting thousands and thousands and thousands of human traces doing it.

Lex Fridman 25:50 But if-

Balaji Srinivasan 25:50 But it is a predator-prey on that. Go ahead.

Lex Fridman 25:52 Well, and then the computer-

Balaji Srinivasan 25:53 Or millions of human traces. I don't know.

Lex Fridman 25:55 The computer just sees the JavaScript. It needs to be able to look outside the simulation for the computer. The world is like... you-

Balaji Srinivasan 26:03 Right.

Lex Fridman 26:03 It doesn't... The computer doesn't know about-

Balaji Srinivasan 26:05 Right.

Lex Fridman 26:05 ... the physical world.

Balaji Srinivasan 26:05 Yeah.

Lex Fridman 26:06 So it has to look outside of its world and introspect back on this simple box-

Balaji Srinivasan 26:11 Right.

Lex Fridman 26:11 ... is, which is kind of, you know... I think that's exactly what mushrooms do, or like psychedelics, is you get to go outside and look back in, and that's-

Balaji Srinivasan 26:19 Yeah, it's interesting.

Lex Fridman 26:19 ... what a computer needs to do.

Balaji Srinivasan 26:20 I, I, you know, I do wonder whether they actually give people insight or whether they give people the illusion of insight. Um-

Lex Fridman 26:26 Is there a difference?

Balaji Srinivasan 26:27 Yeah. Because a- well, actual insight, you know... Actual insight is, again, Maxwell's equations. You're g- able to shift the world with that. There's a lot of practical devices that work. The illusion of insight is, "I'm Jesus Christ," and nothing happens, right? So, I don't know. I think those are quite different.

Lex Fridman 26:46 (laughs) Uh, I don't know. I, I, I think you can fake it till you make it on that one, which is, um... Insight, in some sense, is revealing a truth that was there all along.

Balaji Srinivasan 26:57 Yeah, so, I mean, I guess like I'm talking about technical insight where you, you have... This is the thing, you know, we were talking about actually m- before the podcast, like technical truths versus political truths, right? Some truths, they're, they're on a spectrum, and there's some truths that are actually entirely political in the sense that if you can change the software in enough people's heads, you change the, the value of the truth. For example, the location of a border is effectively consensus between large enough groups of people. Uh, who is the CEO? That's, you know, consensus among a certain group of people. What is the value of a currency or any stock, right? That, that market price is just psychology of a bunch of people. Like literally, if you can change enough people's minds, you can change the value of the border or the position of the hierarchy or the value of the currency. Those are purely political truths. Then all the way on the other end are technical truths that exist independent of whatever any one human or all humans think. Like, uh, the gravitational constant, right? Or the diameter of a virus. Those, those are just... Those exist independent of the human mind. Change enough human minds, it doesn't matter. Those, those remain constant. And, um, and you have things that are interestingly in the middle where cryptocurrency has tried to pull more and more things from the domain of political truths into technical truths, where they say, "Okay, the one social convention we have is, um, that if you hold this amount of Bitcoin or, or that if you hold this private key, you hold this Bitcoin." And then we make that very hard to change because you have to change a lot of technical truths. So y- you can push things to this interesting intermediate zone.

Lex Fridman 28:28 Yeah, the question is how m- how, how much of our world can we push into that?

Balaji Srinivasan 28:33 Right.

Lex Fridman 28:34 And that takes us in a non-linear fascinating (laughs) journey to the question I wanted to ask you in the beginning-

Balaji Srinivasan 28:43 Sure.

Lex Fridman 28:43 ... which is, um, this political world that you mentioned and the world of political truth. As we know it, in the 20th century and the early 21st century, what do you think works well and what is broken about government?

Balaji Srinivasan 28:57 The fundamental thing is that we can't easily and peacefully start new opt-in governments. And-

Lex Fridman 29:06 Like startup governments.

Balaji Srinivasan 29:07 Yeah. And what I mean by that is basically, um, you can start a new company, you can start a new community, you can start a new currency even these days. You don't have to beat the former CEO in a duel to start a new company. Um, you don't have to become head of the World Bank to start a new currency, okay? Um, because of this, yes, if y- you can... If, if you're... if you want to, you can join, I don't know, uh, Microsoft or name some company that's a... GameStop, and you can try to reform it, okay? Or you can start your own. And the fact that both options exist mean that, you know, you've... you can actually just start from scratch, and that's just... I mean, the same reason we have a clean piece of paper, right? I- I mentioned this actually in, in The Network State book. I'll just quote this bit, but we want to be able to start a new state peacefully for the same reason we want a bare plot of Earth, a blank sheet of paper, an empty text buffer, a fresh start, or a clean slate, because we want to build something new without historical constraint, right? For the same reason you hit plus and do docs.new, you know, like create a new doc. It's for the same reason, right? Because you don't want backspace, you don't want to have just like...... 128 bytes of space, 128 kilobytes, and just have to backspace the old document before creating the new one. So, that's a fundamental thing that's wrong with today's governments. And it's a meta point, right? Because it's not any one specific reform, it's a meta-reform of being able to start new countries.

Lex Fridman 30:30 Okay, so that's one problem, but there, you know, you could push back and say that's, that's a f- feature beca- you know, a lot of people argue that tradition is power. Through generation, if you try a thing long enough, which is the way I see marriage, there's value to the struggle and the journey you take through the struggle, and you grow and you develop ideas together, you grow intellectually, philosophically together. And that's the idea of a nations that spans generations, that you have a tradition that becomes more, that, that strives towards the truth and is able to arrive there. Or no, not arrive, but take steps towards there through the generations. So y- you may not want to keep starting new governments. You might want to, uh, stick to the old one and, uh, improve it one step at a time. So just because you're having a fight inside a marriage doesn't mean you should get a divorce and go on Tinder and start dating around. That's the, uh, that's the pushback. So it's-

Balaji Srinivasan 31:31 Sure.

Lex Fridman 31:32 ... it's not obvious that this is a strong feature to have-

Balaji Srinivasan 31:35 Sure.

Lex Fridman 31:35 ... you know, to launch new governments.

Balaji Srinivasan 31:36 There's several different kind of lines of attack, or, or debate, or whatever on this, right? First is, uh, yes, there's obviously value to tradition, and, uh, you know, people say, "This is Lindy and that's Lindy. It's been proven for a long time," and so on. But of course, there's a tension between tradition and innovation, you know? Like, going to the moon wasn't Lindy (laughs) , just it was awesome, and you know, like artificial intelligence is something that's very new. New is good, right? And this is a tension within humanity actually itself, 'cause you know it's way older than all of these nations. I mean, humans are tens of thousands of years old. The ancestors of humans are millions of years old, right? And you go back far enough, and the time that we know today of the sessile farmer and soldier is, if you go back far enough, you wanna be truly traditional, well, we're actually descended from hunter-gatherers who were mobile, and wandered the world, and there weren't borders and so on. They kind of went where they want, right? And you know, people have, you know, had done historical reconstructions of like skeletons and, and stuff like that, and, uh, many folks report that the transition to agriculture and being sessile, um, resulted in, you know, diminution of height, you know, people had like tooth decay and stuff like that. The skeletons, they, uh, people had traded off upside for stability, right? That's what the state was, that was what these sessile kinds of things were. Now, of course, um, they, they had more likelihood of living, uh, consistently, you could support larger population sizes, but it had lower quality of life, right? And so the hunter-gatherer era, you know, maybe that's our, our collective recollection of a Garden of Eden, where people, you know, just like a, a spider kind of knows i- innately how to build webs, or a beaver knows how to build dams, you know, some people theorize that, uh, the entire Garden of Eden is like, um, a sort of built-in neural network recollection of this, you know, pre-sessile era, where we were able to roam around and just pick off fruits and so on, low population density. So, point is that I think what we're seeing is a V3. You go from the hunter-gatherer to the farmer and soldier, the sessile nations are here and they've got borders and so on, to kind of the V3, which is the digital nomad, the new hunter-gatherer. We're going back to the future, because you know what's even older than nations is no nations, right? Even more traditional than tradition is, you know, being international, right?

Lex Fridman 33:57 Mm-hmm.

Balaji Srinivasan 33:58 And so we're actually tapping into that other huge thread in humanity, which is the desire to explore, pioneer, wander, innovate, you know?

Lex Fridman 34:07 So the way-

Balaji Srinivasan 34:07 And I think that's important.

Lex Fridman 34:08 ... to make America great again is to dissolve it completely in- into oblivion. No, that's a joke.

Balaji Srinivasan 34:13 I, I, see yeah, yeah, yeah, I know it's a joke. B- but the thing that... Well-

Lex Fridman 34:16 (laughs) Humor, I'm learning this new thing.

Balaji Srinivasan 34:18 Yes, a new thing for the ro- uh, yeah. You, the chatbot emulation isn't fully working there.

Lex Fridman 34:21 Yeah, yeah, glitch. This, we're in the, in the beta-

Balaji Srinivasan 34:24 And let me just say, say one other thing about this, which is, you know, there are, I mean, everybody in the world to fr- okay, let's say, I don't know what percentage, let's say 99.99%, uh, or it's rounds to that number, of political discourse in the US focuses on trying to fix the system.

Lex Fridman 34:44 Hmm.

Balaji Srinivasan 34:44 If those folks, I mean, 0.01% of the energy is going towards building a new system, that seems like a pretty good portfolio strategy, right? Or 100% are supposed to go and edit this code base from 200-something years ago. I mean, the most American thing in the world is going and, you know, leaving your country in s- in search of a better life. America was founded 200 years ago by the Founding Fathers. It's not just a nation of immigrants, it's a nation of emigrants, right? Emig- emigration, you know, from other countries to the US, and actually also, emigration within the US. There's an amazing YouTube video called, um, it's like, 50 States, US Population, I think 1790 to it says 2050, so they've got a simulation, so you just stop it at 2019 or 2020. But it shows that, like Virginia was like number one early on, and then it lost ground and like New York gained.

Lex Fridman 35:29 Mm-hmm.

Balaji Srinivasan 35:29 And then like Ohio was a big deal in the early 1800s, and it was like father of presidents and generally all these presidents, and later Illinois and Indiana, and then California only really came up in the, the 20th century, like during the Great Depression. And now we're entering the modern era, where like Florida and Texas have risen and New York and California have dropped.

Lex Fridman 35:46 Mm-hmm.

Balaji Srinivasan 35:46 And so interstate competition, it's actually just like inter-currency competition, you know, you've got trading pairs, right? You, you know, sell BTC, buy ETH. You sell, you know, eh, eh, Solana or ZC-, you know, sell, eh, Monero, buy Zcash, right? Each of those trading pairs gives you signal for today on this currency is down or up relative to this other currency. In the same way, each of those migration pairs, someone goes from New York to Ohio, Ohio to California, gives you information on the desirability of different states. You can literally form a pairs matrix like this over time.

Lex Fridman 36:19 Mm-hmm.

Balaji Srinivasan 36:19 Uh, very much like the link matrix. That's shaped America in a huge way, and so, you know, you, you ask, A, if th- if this nation of emigrants that was founded by men younger than us, by the way, the Founding Fathers were often in their 20s, right? Who, um, you know, endorsed the concept of a proposition nation, who've given rise to a country of founders and, and pioneers, uh, who've literally gone to the moon, right? Um, those folks would think that this is the end of history?... that, that's it? We're done? Like, we've, we've done everything else? I mean, there's people in technology who believe, and I agree with them, that we can go to Mars, that we might even be able to end death. But we can't innovate on something that was 230 years old? You know?

Lex Fridman 36:59 So there is a balance certainly to, to strike. Uh, the, the American experiment is fascinating, nevertheless. So one argument you could make is actually that we're in the very early days of this V2. If th-

Balaji Srinivasan 37:13 Yeah.

Lex Fridman 37:13 So what you describe as V2, you could make the case that we're not ready for V3. That we're just actually, uh, trying to figure out the V2 thing. You're trying to, like-

Balaji Srinivasan 37:23 Hmm.

Lex Fridman 37:24 ... skip-

Balaji Srinivasan 37:24 When are we ever ready?

Lex Fridman 37:26 Now again, we'll go back to marriage, I think, uh, and, and having kids kind of thing. I think everyone who has kids is never really ready to be kids. That's the whole point, you dive in. Okay. But, uh, I mean, you, you mentioned that you Is there other criticisms of government that you can provide as we know it today, before we kind of-

Balaji Srinivasan 37:44 Sure. S-

Lex Fridman 37:44 ... outline the ideas of, of V3. Let's stick to V2.

Balaji Srinivasan 37:48 I'll, I'll give a few, right? And so a lot of this stuff will go into the version ... So I've got, you know, this book The Network State, um, which, which covers some of these topics.

Lex Fridman 37:56 Does Network State have a subtitle?

Balaji Srinivasan 37:58 Um, it, it is, uh, The Network State: How to Start a New Country.

Lex Fridman 38:02 How to Start a New Country.

Balaji Srinivasan 38:03 And ... But, um, I just have it at thenetworkstate.com.

Lex Fridman 38:05 I should say-

Balaji Srinivasan 38:06 Yeah.

Lex Fridman 38:06 ... it's an excellent book that you should get. I read it on Kindle, but there's also a website. And, uh, Balaji said that he's constantly working on improving it, changing it. By, b- by the, by the time the whole project is over, it'll be a different book than it was-

Balaji Srinivasan 38:24 Yeah.

Lex Fridman 38:25 ... in the beginning.

Balaji Srinivasan 38:25 I think so.

Lex Fridman 38:26 It's always shedding its old skin.

Balaji Srinivasan 38:29 Well, I, I, I wanted to get something out there and get feedback, and, and whatnot. Just like an app, right? You know, you ... Again, you have these two poles of an app is highly dynamic and you're, you're accustomed to having updates all the time, and a book is supposed to be static. And there's a value in something static, something unchanging and, and so on. But in this case, I'm glad I kind of shipped a version 1.0 and, uh, you know, the, the next version, um, you know, I'm gonna split it into like, uh, tentatively motivation, theory, and practice. Like motivation, like what is the sort of political philosophy and so on that motivates me at least through this, which you can take or leave, right? And then theory as to why Network State is now possible, and I can define it in a second. And then the practice is zillions of practical details and everything from roads to diplomatic recognition and so on. Um, funding, founding, all that stuff. A lot of the stuff actually I left out of V1 simply because I wanted to kind of get the desirability of it on the table and then talk about the, the feasibility and the-

Lex Fridman 39:26 I should actually linger on that briefly in t- in terms of things we can revolutionize. Like, um, one of the biggest innovations I think that Tesla does with the way they think about the car, with the deploy of the car, is not the automation or the electric to me. It's the, uh, over the air updates.

Balaji Srinivasan 39:44 Mm-hmm. Right.

Lex Fridman 39:45 Be able to send instantaneously, uh, updates to the software that completely changes the behavior, the UX, everything about the car. And so I do think it would be interesting 'cause books are a representation of human knowledge, a, um, a snapshot of human knowledge. And it would be interesting that we ... If we can somehow figure out a system that allows you to do sort of like a GitHub for books. Like if I buy a book on Amazon, without having to pay again, can I get updates? Like V point ... Uh, V1.1, V1.2, and there's like release notes.

Balaji Srinivasan 40:25 Right.

Lex Fridman 40:25 That, that would be, that would be incredible. It's not, not enough to do like a second edition or a third edition, but like minor updates that's not just on your website but actually go into the, the model that we use to buy books.

Balaji Srinivasan 40:40 Right.

Lex Fridman 40:40 So I spend my money, maybe I, I'll do a subscription service for five bucks a month where I get regular updates to the books.

Balaji Srinivasan 40:48 Right.

Lex Fridman 40:48 And then there's an incentive for authors to actually update their books such that it makes sense for the subscription. And then that means your book isn't just a snapshot but is a lifelong project.

Balaji Srinivasan 40:59 Right. So-

Lex Fridman 41:00 If you care enough about the book.

Balaji Srinivasan 41:01 So I think there's a lot that can be done there, because actually in going through this process in many ways the most traditional thing I did was self-publish ebook on Kindle.

Lex Fridman 41:10 Mm-hmm.

Balaji Srinivasan 41:10 Right? Uh, why? Because basically like, you know, i- if you actually ink a deal with a book publisher, first they, you know, they'll give you some advance. I, I didn't need the advance or anything. But second is all these constraints. Um, oh, you know, you want to translate into this or you want to do this other format or we want to update it, you have to go and now talk to this other party, right? You know, and, uh, also the, the narrowing window of what they'll actually publish, it gets narrower and narrower. You can see all these, you know, meltdowns over young adult novels and stuff on Twitter, but it's, it's, it's more than that. So, you know, actually having an Amazon page, it's just like a marker that a book exists.

Lex Fridman 41:46 Mm-hmm.

Balaji Srinivasan 41:46 Okay? And, uh, now I've got an entry point where if someone says, "Okay, I like this tweet but how do I kind of get the, um ..." That, that might be a, a concept from like the middle of chapter three, right? How do I get the thing from front to back? I can just point them at thenetworkstate.com. That is, import this, right, this one entry point, okay? And, um, you mentioned like subscription and, and, and money and so on and so forth, and I think people are paying for content online now with newsletters and so on. But I've chosen to and I will always have the thing free, um, and I want it on ... Or, uh, you can get the Kindle version on Amazon simply because they ... You have to kind of set a price for that. But thenetworkstate.com, what I want to do is have that optimized for every Android phone so people in India or Latin America or Nigeria can just tap and open it. Gonna do translations and stuff like that. Uh, Greg Foder of AltspaceVR, you know, founder of AltspaceVR, he, you know, he, he sold that and, uh, he coded the website. And, um, you know, I worked with him on it, and there's another designer who, um-... Elijah, and, uh, it was basically just a three-person group. And we thought we had something pretty nice. But one thing I was really, uh, pleasantly surprised by is how many people got in touch with us afterwards and asked us if we could open source the software to create this, this website, right? Because it's actually a ... Y- you can try it on mobile. I think it's actually, um, in some ways, a better experience than Kindle.

Lex Fridman 43:07 Mm-hmm.

Balaji Srinivasan 43:07 And, uh, so that was interesting because, um, I do think of the website as like a V1 version of, uh, this concept of a book app, right? For example, imagine if you have the Bible and the Ten Commandments aren't just text, but there's like a checklist and there's a gateway to a Christian community there, and, um, you know, the practice is embedded into the thing, you know? Like, if, if ... Do you know brilliant.org?

Lex Fridman 43:31 Mm-hmm.

Balaji Srinivasan 43:31 Amazing site. I love this site. Brilliant is basically mobile-friendly tutorials, and you can kind of just swipe through, you know, you're in line at Starbucks or, you know, getting on a plane or something, you just swipe through and you just get really nice micro lessons on things, and it's just interactive enough that your brain is working and you're problem-solving. And, uh, sometimes you'll need a little pen and paper. But that format, um, of sort of very mobile-friendly, just continuous learning, I, I, I'd like to do a lot more with that. And so that's kind of where we're gonna go with the, the book app.

Lex Fridman 44:05 So the, uh, uh, there's a lot of fun stuff about the way you did-

Balaji Srinivasan 44:08 Sure.

Lex Fridman 44:08 ... at least V1 of the book, which is, you have like a one-sentence summary, one-paragraph summary, TLDR, and like one-image summary.

Balaji Srinivasan 44:20 Yeah.

Lex Fridman 44:20 Which is, um ... I think, honestly, it's not even about a short attention span, it's a really good exercise about summarization, condensation, and really-

Balaji Srinivasan 44:30 Yeah.

Lex Fridman 44:30 ... like helping you think through what is the key insight. Like we mentioned the, the prime number maze that reveals something central to the human condition, which is struggling against the limitation of our, of our minds, and in that same way, you summarize the network state in the book. So

Lex Fridman 44:50 let, let, let's actually jump right there and let me ask you, what is the network state?

Balaji Srinivasan 44:55 What is the network state? So I'll give it in a sentence and I'll also give it in an image, right? So the informal sentence. A network state is a highly aligned online community with a capacity for collective action that crowdfunds territory around the world and eventually gains diplomatic recognition from preexisting states. Okay? So just taking those pieces. Highly aligned online community, that is not Facebook, that is not Twitter. F- people don't think of themselves as Facebookers or Twitterians, right? That's just a collection of hundreds of millions of people who just fight each other (laughs) all day, right? It's a fight club. A company is highly aligned where, you know, you'll put a task into the company Slack, and on the, r- if you do it in all-hands, about 100% of, of the people in the company Slack will do it, so they're highly aligned in that way. But online communities don't tend to be highly aligned. Online communities tend to be like a Game of Thrones fan club or something like that.

Lex Fridman 45:42 Mm-hmm.

Balaji Srinivasan 45:43 Where, you know, on, on a Twitter account, you might get .1% of people engaging with something. It's not to 100%. If you combine the degree of alignment of a company with the scale of a community, that's like what a highly aligned, you know, online community is, right? So to get 1,000 or 10,000 people who can collectively do something as simple as just all liking something on Twitter, for example. Why would they do that? They're a, they're a guild of electrical engineers, they're a guild of graphic designers, and you've got 1,000 people in this guild, and every day, somebody is asking a favor from the guild, and the other 999 people are helping them out. For example, "I've just launched a new project," or, "I'd like to get a new job. Can somebody help me?" And so on. And so you kind of give to get. You're, you know, you're helping other people in the community and you're kind of building up karma this way, and then sometimes you spend it down. Like Stack Overflow has this karma economy. It's not meant to be an internal economy that is, um, like making tons and tons of money off of ... It's sort of to keep score, right? That's the highly aligned online community part. Then capacity for collective action, I just kind of described that, which is, at a minimum, you, you, you don't have a highly aligned online community unless you have 1,000 people and you paste in a tweet and 1,000 of them RT it, or, or like it, okay?

Lex Fridman 46:52 Mm-hmm.

Balaji Srinivasan 46:52 If you can't even get that, you don't have something. If you do have that, you have the basis for at least collective digital action on something, okay? And you can think of this as a group of activists. You can think of it as j- for example, let's say, I mentioned a guild, but let's say they're a group that wants to raise awareness, uh, of the fact that life extension is possible, right? Every day, there is a new, um, tweet on, I don't know, whether it's, uh, metformin research or Sinclair's work, or David Sinclair, right? Andrew Huberman has good stuff here, you know, or, um, there's a LongevityVC, there's a bunch of folks working this area. Every day, there's something there, and literally, the purpose of this online community is ra- raise awareness of longevity, and of the 1,000 people, 970 go and like that. That's pretty good, right? That's solid. You've got something there. You've got, you've got a laser, right? You've got something which you can f- focus on something, because most of the Web 2.0 internet is entropic. You go to Hacker News, you go to Reddit, you go to Twitter, and you're immediately struck by the fact that it's like 30 random things. Random. It's just a box of chocolates. It's meant to be, you know, w- we're-

Lex Fridman 47:56 Some of them look delicious.

Balaji Srinivasan 47:57 Some of them look delicious. Novelty, we can over-consume novelty, right? So, you know, we were talking about it earlier, the balance between tradition and innovation, right? Here is a different version of that, which is, um, entropy going in a ton of different directions due to novelty, versus, uh, like focus, you know? It's like, it's like heat versus work, you know? Heat is entropic, and work force along a distance, you're going in a, in a direction, right? And so, if those 30 links on, you know, the next version of Hacker News or Reddit or something like so brilliant is just, that's leveling you up. The, the 30 things you click, you've just gained a skill as a function of that, right? So these kinds of online communities, I don't know what they look like. They probably don't look like the current social media. They l- they... Just like, for example, I know this is a meta analogy, but in the 2000s, people thought Facebook for Work would look like Facebook. And, you know, David Sacks, you know, founded and sold the company Yammer that was partially on that basis. It was, it was fine. It was a billion-dollar company. But Facebook for Work tended ... wh- was actually Slack.... right?

Lex Fridman 48:55 Mm-hmm.

Balaji Srinivasan 48:55 It looked different. It was more chat-focused, it was less image-focused and whatnot. What does the platform for a highly aligned online community look like? I think Discord is the transitional state, but it's not the end state. Discord is sort of chatty. The work isn't done in Discord itself, right? The cryptocurrency for tracking, or the cryptokarma for sort of tracking people's contributions is not really done in Discord itself. Discord was not built for that.

Lex Fridman 49:19 Mm-hmm.

Balaji Srinivasan 49:19 And I don't know what that UX looks like. Maybe it looks like tasks, uh, m- you know, like, uh, may- maybe it looks, uh, something different. Okay.

Lex Fridman 49:26 So wait, wait, wait. Le- le- let me linger on this.

Balaji Srinivasan 49:28 Sure.

Lex Fridman 49:28 So, you're actually, uh... there's some people who might not be even familiar with Discord or Slack or, and so on.

Balaji Srinivasan 49:34 Mm-hmm.

Lex Fridman 49:34 Even these platforms have, like, communities associated with them.

Balaji Srinivasan 49:39 Yes.

Lex Fridman 49:40 Meaning, meaning the, the big, the, like the meta-community of people who are aware of the feature set and that you can do a thing. That this is a thing, and then you can do a thing with it. Discord, like when I first realized it, I think it was borne out of the gaming world.

Balaji Srinivasan 49:54 Yes.

Lex Fridman 49:55 It's like, holy shit, this is like a thing. There's a lot of people that use this.

Balaji Srinivasan 50:01 Right.

Lex Fridman 50:01 There's also a culture that's very difficult to escape that's associated with Discord that spans all the different communities within Discord.

Balaji Srinivasan 50:09 Right.

Lex Fridman 50:09 Reddit is the same. Even though there's different subreddits, there's still, because of the migration phenomenon maybe, there's still a culture to Reddit and so on.

Balaji Srinivasan 50:19 Yes.

Lex Fridman 50:20 So the... I- I- I'd like to sort of try to dig in and understand what's the difference between the online communities that are formed and the platforms on which those communities are formed?

Balaji Srinivasan 50:32 Sure.

Lex Fridman 50:32 Really important.

Balaji Srinivasan 50:33 Yes, it is. It is. So for example, an office. A good design for an office is frequently you have, um, you know, the, uh, the commons, which is, like, the lunchroom or the gathering area. Then everybody else has a cave on the border they can kind of retreat to.

Lex Fridman 50:47 Cave and the commons. I love it. By the way, I was laughing internally about the heat versus work. I think that's gonna stick with me. That's such an interesting way to see Twitter.

Balaji Srinivasan 50:55 Yeah.

Lex Fridman 50:56 Like, is this heat, or is this, is this thread like g- 'cause there's a lot of stuff going on.

Balaji Srinivasan 51:03 Right.

Lex Fridman 51:04 Is it just heat, or are we doing some, like... Is it, is there a directed thing-

Balaji Srinivasan 51:07 Exactly.

Lex Fridman 51:07 ... that's gonna be productive at the end of the day?

Balaji Srinivasan 51:09 That's right.

Lex Fridman 51:09 I love this. I'm never see d- diff-... I mean, anyway. The cave and the commons is, uh, is a really nice, uh... So that has to do with the layout of an office that's effective.

Balaji Srinivasan 51:18 Yep. That's right. And, uh, so you can think of many kinds of social networks, um, as being on the cave and commons continuum. For example, Twitter is just all commons. The caves are just like individual DMs or DM threads or whatever, but it's really basically just one gigantic global public fight club, for the most part, right? Then you have-

Lex Fridman 51:39 Or love club.

Balaji Srinivasan 51:40 Well, somewhat love, but mostly fight. Or actually it's-

Lex Fridman 51:43 I love aggressively, that's all.

Balaji Srinivasan 51:44 Yeah. I mean, the, the way I think... I mean, Twitter is like a cross between a, you know, a, a library and a civil war, you know? It's something where you can learn and, but, but you can also fight if you choose to fight, right? Um-

Lex Fridman 51:59 Yeah. Well, I mean, it's um, uh, because of its, the commons structure of it, it's a mechanism for virality of anything.

Balaji Srinivasan 52:10 Yeah. So-

Lex Fridman 52:10 You just described the kind of things that become viral.

Balaji Srinivasan 52:13 Yeah. Meaning no, no offense to librarians. It's like a library and Liberia. Liberia was racked by civil war for, for many years, right? Um-

Lex Fridman 52:20 Uh, libraries are one of my favorite sets, uh, for porn. Just kidding. Jokes. I'm learning, as that's probably crossing the line, uh, for the engineers working on this humor module. Maybe take that down a notch.

Balaji Srinivasan 52:32 Yeah. Um, gosh, uh, we were just talking about-

Lex Fridman 52:35 Uh, what-

Balaji Srinivasan 52:35 Oh yeah, so continue. Go ahead, continue. Right?

Lex Fridman 52:36 Uh, it's, Twitter's a commons.

Balaji Srinivasan 52:37 Yeah. So Twitter's a commons, and Facebook is like... It's got all these warrens and stuff. Facebook, it's very difficult to reason about, uh, like privacy on that. And the reason is, I think it's easy to understand when something is completely public, like Twitter, or completely private, like Signal. And those are the only two modes, I think, in which one can really operate. When something is quasi-private, like Facebook, you have to just kind of assume it's public. Because if, if it's interesting enough, it'll go outside your friend network, and it'll get screenshotted or whatever and posted. And so, you know, Facebook is sort of, sort of forced into default public, despite its privacy settings. Uh, you know, for anybody who says something interesting. Uh, you know, if it's like, uh, y- you can figure out all their dials and stuff like that, but just hard to understand unless it's totally private or totally public, right? You have to basically treat it if it's totally public if it's not totally private. Okay. At least under a real name. I'll come back to pseudonym. So you've got Twitter, that's, that's total commons. Facebook, which is like a warrens, you know, is like, you know, it's like rabbit warrens or like a ant colony where you don't know where information is traveling. Then you've got Reddit, which has sort of your global Reddit and then all the subreddits. That's a different model of cave and commons. I think one of the reasons it works is that you have individual moderators where something is totally off-topic and acceptable, uh, totally off-topic and unacceptable in this subreddit and totally on-topic and acceptable in another. That's like kind of a, you know, precursor of the digital societies I think that we're gonna see, that actually are become physical societies. Like, lots and lots of subreddit-like things, you know, uh, become physical societies. Then you start going further into, like, Discord, where it's, it's more full-featured than, um... You know, as, as you go Twitter, Facebook, Reddit, now you jump into Discord, and Discord is a bunch of individual communities that are connected, and you can easily sort of jump between them, right? And then you have Slack, and you know, yes, you can use, um, Slack to go between different company Slacks, but Slack historically at least, I'm not sure what their current policy is, historically they discouraged public Slacks. So there, it's mostly like you have your main Slack for your company, and then you sometimes may jump into like a, let's say you've got a design consultant or somebody like that, you'll jump into their Slack. But Discord is, you've got way more Discords usually that you jump into than, than Slacks, right? Okay.

Lex Fridman 54:51 Well, yeah, and let me ask you then on that point,

Lex Fridman 54:54 because there is a culture, one of th-... things I discovered on Reddit and Discord of anonymity or s- pseudonyms or usernames-

Balaji Srinivasan 55:02 Yes.

Lex Fridman 55:02 ... that don't represent the actual name. On Slack is an example of one, because I think I did a, I used to have a Slack for like deep learning course that I was teaching, and that was like a h- very large, like 20,000 people, whatever. But so you could grow quite large, but there was a culture of like, "I'm going to represent my actual identity, my actual name."

Balaji Srinivasan 55:21 That's right.

Lex Fridman 55:21 And then the same stuff on Discord. I think I was the only asshole using my actual name on there.

Balaji Srinivasan 55:26 Hmm.

Lex Fridman 55:26 It's like everybody was using, uh, pseudonyms. So what's, what's the role of that in the online community?

Balaji Srinivasan 55:33 Well, so I actually gave a talk on this a few years ago called, uh, the Pseudonymous Economy. Okay. And, um, it's come about faster than I expected, uh, but I did think it was gonna come about fairly fast. And essentially the concept is obviously we've had ... So first, anonym, pseudonym, real name, right?

Lex Fridman 55:49 Can you describe the difference between that?

Balaji Srinivasan 55:50 Difference? Anonymous is like 4chan, uh, where there's no tracking of a name. You're, you know, there's zero reputation associated with an identity, right? Pseudonymous is like much of Reddit, where there's a persistent username and it has karma over time, but it's not linked to a, the global identifier that is your state name, all right?

Lex Fridman 56:11 Mm-hmm.

Balaji Srinivasan 56:11 So your, quote, "real name," even the term real name, by the way, is a misnomer because it's like your Social Security name, like Social Security number. It's your official government name. It's your, it's your state name. It is the tracking device. It's a, it's a air tag that's put on you, right? Why do I say that, right? Another word for a name is a handle. And so just visualize like a giant file cabinet. There's a handle with Lex Fridman on it that anybody, the billions of people around the world can go up to and they can pull this file on you out. Images of you, things you said, like billions of people can stalk billions of other people now. That's a very new thing. And I actually think this will be a transitional era in like human history. We're actually gonna go back into a much more encrypted world. And-

Lex Fridman 56:54 O- O- O- Okay, let me linger on that because, um-

Balaji Srinivasan 56:57 Sure.

Lex Fridman 56:57 ... another way to see real names is the label on a thing that can be canceled.

Balaji Srinivasan 57:04 Yes, that's right. A- In fact, there's a book called Seeing Like a State which actually talks about the origins of surnames and whatnot. Like if you have a guy who is that guy with brown hair. That's like an analog identifier. It could mean 10 different people in a village.

Lex Fridman 57:19 Mm-hmm.

Balaji Srinivasan 57:19 But if you have a first name, last name, okay, that guy can now be conscripted. You can go down with a list, a list of digital identifiers, pull that guy out, pull him, you know, into the military for conscription, right?

Lex Fridman 57:29 Mm-hmm.

Balaji Srinivasan 57:30 So that was like one of the purposes of names, was to make masses of humans legible to a state, right?

Lex Fridman 57:37 Mm-hmm.

Balaji Srinivasan 57:37 Hence, seeing like a state. You can see them now, right? See, digital identifiers, uh, one thing that people don't usually think about is pseudonymity is itself a form of decentralization. So, you know, people know Satoshi Nakamoto was pseudonymous. They also knew his inter- decentralization. But one way of thinking about it is, let's say his real name, okay, or his state name is Anode, okay? Attached to that is, uh, every database, you know, his, uh, his Gmail, his, you know, Facebook if he had one, every government record on him, right? All of these databases have that state name as the foreign key, right? Um, and so it can go and look things up in all of those, uh, databases, right?

Lex Fridman 58:24 Mm-hmm.

Balaji Srinivasan 58:24 And so it's that, like think of it as being the center of a giant network of all of these things. When you go and create a pseudonym, you're budding off a totally new node that's far away from all the rest. And now he's choosing to attach Bitcoin Talk and bitcoin.org and the GPT signatures of, of the code if he ch- choose to do that. All those things, the digital signatures are all attached to this new decentralized name because he's instantiating it, not the government, right? One way of thinking about it is the root administrator of the, quote, "real name" system is the state, because you cannot simply edit your name there, right? You can't just go, you can't log into usa.gov and backspace your name and change it. Moreover, um, your birth certificate, all this stuff, that's fixed and immutable, right? Whereas you- we take for granted that on every site you go to, you can backspace, you can be like, "Call me Ishmael," you know, walk into a site, u- use whatever name you want. You just have to use the same name across multiple sites. You can do that. And if not, you don't have to. One thing that we're seeing now actually is at the level of kids, you know, the younger generation, um, Eric Schmidt several years ago mentioned that, you know, people would like change their names when they became adults, so that they could do that. This is kind of already happening. People are using, uh, I've remarked on this many years ago, search resistant identities.

Lex Fridman 59:39 Mm-hmm.

Balaji Srinivasan 59:39 Okay? Why? They have their Finsta, which is their, quote, "fake name" Instagram, and Rinsta, which is their real name Instagram.

Lex Fridman 59:46 Oh, this is cool.

Balaji Srinivasan 59:46 Okay? And what's interesting is on their Rinsta, they're their fake self because they're in their Sunday best and, you know, smiling. And this is the one that's meant to be search indexed, right?

Lex Fridman 59:56 Mm-hmm.

Balaji Srinivasan 59:56 On the Finsta, with their fake name, this is just shared with their closest friends. They're their real self and they're, you know, hanging out at parties or whatever, you know. And so this way they've got something which is the public persona and the private persona, right? The public persona that's search indexed and the private persona that is private for friends, right? And so organically, people are... You know, like Jeanne Jacobs, she talks about like cities and how, you know, they're organic and, and whatnot. Like, uh, some of the mid-20th century guys, the architecture they had removed shade from, uh, you know, like, like awnings and stuff like that got removed. So this is like the restoration of like awnings and shade and structure so that you're not always exposed to the all-seeing web crawler, that eye of Sauron, which is like Googlebot just indexing everything. These are search resistant identities in that like eye just sort of passes over you, like, you know, in the Terminator, like in the Terminator eye just kind of passes over you, right? So search resistant identity is not pulled up, it's not indexed, right? And now you can be your real self.... and so, we've had this kinda thing for a while with communication. The new thing is that cryptocurrency has allowed us to do it for transaction, hence the pseudonymous economy, right? And, um, as you go from anonymous, pseudonymous, real name, these each have their different purposes. But the new concept is that pseudonym, you can have multiple of them, by the way. Your, your ENS name, you could have it under your, quote, " real name" or state name, like lexfriedman.eth, but you could also be, uh, punk6529.eth, okay?

Lex Fridman 1:01:26 Mm-hmm.

Balaji Srinivasan 1:01:26 And now you can earn, you can sign documents, you can boot up stuff, you can have a persistent identity here, okay, which has a level of indirection to your real name. Why is that very helpful? Because now it's harder to both discriminate against you and cancel you. Concerns of various factions are actually obviated, or at least partially addressed, by going pseudonymous as default, right? It is the opposite of bringing your whole self to work, it's bringing only your necessary self to work, right? Only show those credentials that you need, right? Now, of course, you know, anybody who's into cryptocurrency understands Satoshi Nakamoto and so on is for this, but actually many progressives are for this as well. Why? You know ban the boxes? It's like, you're not supposed to ask about, like, felony convictions when somebody is, you know, being hired, because they've paid, they've served their time, right? Or, um, you're not supposed to ask about immigration status or marital status in an interview, um, and, uh, you know, people have this concept of blind auditions where, you know, uh, i- if a woman is auditioning for, for a, uh, like a violin seat, they put it behind a curtain so they can't downgrade, you know, her for playing. So, her performance is judged on the merits of its audible quality, not in terms of who, who this, this person is, so this way they don't discriminate versus ma- male or female for, um, for who's, uh, you know, getting a, a violin position. So, you, you combine those concepts, like ban the box, not asking these various questions, blind auditions, and- and then also the concept of implicit bias. Like if- if you, if you, uh, you know, believe this research, people are, uh, unconsciously biased towards other folks, right? Okay, so you take all that, you take Satoshi, and you put it together, and you say, "Okay, let's use pseudonyms." That actually takes unconscious bias even off the table, right? Because, um, now you have genuine global equality of opportunity. Moreover, you have all these people, billions of people around the world that might speak with accents but they type without them, and now if they're pseudonymous, you aren't discriminating against them, right? Moreover, with AI, very soon the AI version of Zoom, you'll be able to be whoever you want to be and speak in whatever voice you want to speak in, right? And, um, you'll be... and that'll happen in real time.

Lex Fridman 1:03:34 Uh, so, I mean, this is really interesting. With, uh... for Finsta and Rinsta, there's some sense in which the fake Instagram, you're saying, is where you could be your real self. Well, my question is, under, uh, under a pseudonym or when you're completely anonymous, is there some sense where you're not actually being your real self, that as a social e- entity, th- like, it is... Human beings are fundamentally social creatures-

Balaji Srinivasan 1:04:10 Mm-hmm.

Lex Fridman 1:04:10 ... and for us to be social creatures, there is some sense in which we have to have a consistent identity that can't be canceled, that can be criticized, uh, or applauded in society-

Balaji Srinivasan 1:04:25 Right.

Lex Fridman 1:04:25 ... and that identity persists through time. So, is there some sense in which we would not be our full, beautiful human selves unless we have a lifelong consistent real name at- attached to us in a digital world?

Balaji Srinivasan 1:04:40 So, this is a complicated topic, but let me make a few remarks. First is, real names, quote-unquote "state names," were not built for the internet. They've actually-

Lex Fridman 1:04:48 State names? (laughs)

Balaji Srinivasan 1:04:49 State names, right? It's actually a way, great way of thinking about it. It's Social Security name, right?

Lex Fridman 1:04:52 Yeah.

Balaji Srinivasan 1:04:52 Um, so your state name, your official name was not built for the internet. Why? They give both too much information and too little. Okay? So too much information, because, um, someone with your name can find out all kinds of stuff about you. Like, for example, um, if someone doesn't want to be stalked, right, their real name is out there, their stalker knows it, they can find address information, all this other kind of stuff, right? Um, and with all these hacks that are happening... Just every day we see another hack, massive hack, et cetera. Uh, that real name can be indexed into data that was supposed to be private, right? Like, for example, you know, the Office of Personnel Management, like the government, the US government, many governments actually, are like a combination of this, of the, uh, surveillance state and the Keystone Cops, right? Why? They, they slurp up all the information and then they can't secure it, so it leaks out the back door. Okay? They've, they basically have, you know, 100 million records of all this very s- 300 million records, all this very sensitive data they just get owned, hacked over and over again, right? And so really, there should be something which totally inverts the entire concept of KYC and what have you, um, and of course comply with the regulations as they are currently written. But also, you should argue privacy over KYC. The government should not be able to collect what it can't secure. It's slurping up all this information, it's completely unable to secure it, it's hacked over and over again. You know, China probably has the entire OPM file, and it's not just that, like Texas is hacked, um, a- and some of these hacks are not even detected yet, right? And these are just the ones that have been admitted. And so, you know, what happens is criminals can just run this stuff and find, you know, oh, okay, so that guy who's got that net worth online and he merges various databases, they've got a bunch of addresses to go and hit, okay? So, in that sense real names were not, state names were not built for the internet, they just give up too much information. In our actually existing internet environment they give up too much information. On the other hand, they also give too little. Why? If instead you give out l- uh, lexfriedman.eth, okay, or a similar crypto domain name or Urbit name or something like that, now that's actually more like a DNS, okay? First, if you've got a d- uh, lexfriedman.eth, what can you do that?... um, some you can do today, some you'll soon be able to do. Uh, you can pay lexstream.eth. You can message lexstream.eth. You can, um, look it up like a social profile. You can send files to it. You can upload and download. Basically, it combines aspects of an email address, um, a, uh, uh, a website, a username, et cetera, et cetera, where you, you know, eventually I think you'll go from email to phone number to ENS address or something like that as the primary online identifier, because this is actually a programmable name, right? Whereas a state name is not, you know, think about it, like a state name will have apostrophes perhaps in it or is that your middle name or this and that. That was a format that was developed for the paper world, right? Whereas the ENS name is developed for the online world. Now, reason to say ENS or something like it, you know, somebody had a, in a village they'll, their name might be Smith because they were a blacksmith or Potter because they were a potter, right? And the same way I think your surname, right now for many people it's dot eth and that reflects the Ethereum community. Your surname online will carry information about you, like dot sol says something different about you.

Lex Fridman 1:08:08 Mm-hmm.

Balaji Srinivasan 1:08:09 Dot btc says yet something different. I think we're going to have a massive fractionation of this over time. We're still in the very earliest days of our internet civilization, right?

Lex Fridman 1:08:17 Mm-hmm.

Balaji Srinivasan 1:08:18 100, 200 years from now, those surnames may be as informative as, say, Chen or Friedman or Srinivasan in terms of what information they carry because the protocol is the civilization fundamentally that you're associated with, right?

Lex Fridman 1:08:28 Right. So the- there's some improvements to the real name that you could do in the digital world. But do you think there's value of having a name that's persist throughout your whole life that is shared between all the different digital and physical communities?

Balaji Srinivasan 1:08:45 I think you should be able to, I think you should be able to opt into that, right?

Lex Fridman 1:08:47 At which stage?

Balaji Srinivasan 1:08:48 At which level? Uh, in terms of the society that you're joining.

Lex Fridman 1:08:51 Wait a minute. So can I murder a bunch of people in Society 1 and then go to Society 2 and be like, "I'm murder free. My name is-"

Balaji Srinivasan 1:09:02 No, no. I don't mean it like that. No, no. Yeah.

Lex Fridman 1:09:03 Which, but, you know-

Balaji Srinivasan 1:09:05 So, which is why that wouldn't work.

Lex Fridman 1:09:05 That's the application I'm interested in

Narrator 1:09:07 (laughs)

Balaji Srinivasan 1:09:07 Okay. Well, I'm not interested in the murder application, but, but, but take-

Lex Fridman 1:09:10 So what, what resists-

Balaji Srinivasan 1:09:10 No, no, I'm just kidding.

Lex Fridman 1:09:11 I would like you to pre-

Balaji Srinivasan 1:09:12 To take that thought.

Lex Fridman 1:09:13 ... to prevent me-

Balaji Srinivasan 1:09:13 Sure.

Lex Fridman 1:09:14 A person who's clearly bad for society from doing that.

Balaji Srinivasan 1:09:17 Sure, sure. Murder is going to be against the rules in almost every society. And I mean, basic- I mean people will argue-

Lex Fridman 1:09:24 Most likely, yeah.

Balaji Srinivasan 1:09:24 Yeah, most likely, right? And, um, the reason I'm thinking of-

Lex Fridman 1:09:27 Except animals.

Balaji Srinivasan 1:09:28 Well, I'm thinking of like the Aztecs or the Mayas or, you know, something like that. There's various, you know, Soviet Union, there's weird, there's weird edgeca- not edge case, unfortunately.

Lex Fridman 1:09:36 Wrong words.

Balaji Srinivasan 1:09:36 Yeah, there's societies unfortunately that have actually, that's why, that's why I asterisked it.

Lex Fridman 1:09:40 Right.

Balaji Srinivasan 1:09:40 But let's say murder is, uh, something that Society 1 probably has effectively a social smart contract or a social contract that says that's illegal, therefore you're in jail, therefore you're deprived of the right to exit. But upon entry into that society, in theory you would have said, "Okay, I accept this," quote, "social contract," right? Obviously if I kill somebody I can't leave, okay? So you've, you've accepted upon crossing the border into there, right? Now, um, as I mentioned, you know, uh, like what is murder? Like people will... I mean there's an obvious answer but as I said there's been human sacrifice in some societies, communism they killed lots of people, Nazism they killed lots of people. Unfortunately there's quite a lot of societies, you know, I wanted to say it's an edge case but maybe many of the 20th soci- century societies around the world have institutionalized some kind of murder, whether it was the Red Terror, you know, in the Soviet Union, or obviously the Holocaust or, you know, the Cultural Revolution or, or Year Zero and, and so on and so forth, right? So my point there is that who's committing all those murders? It was the state, it was the organization that one is implicitly trusting them to track you, right? And how did they commit those murders? Well, how did, how did Lenin, you know... Uh, you know the hanging order? You know what I'm talking about, the hanging order for the kulaks?

Lex Fridman 1:10:56 Yes.

Balaji Srinivasan 1:10:56 Okay. The famous hanging order which actually showed they were actually bloodthirsty. The key thing was he said here's a list of all the, quote, "rich men," the kulaks, go and kill them. The real names, the state names were what facilitated the murder, didn't prevent the murderers there.

Lex Fridman 1:11:10 Mm-hmm.

Balaji Srinivasan 1:11:10 Right? So my point is just in the ethical weighting of it, it's a two-sided thing, right? You're right that the tracking can, you know, prevent disorganized murders but the tracking facilitates unfortunately organized murders. Lists of undesirables were the primary tool of all of these oppressive states in the 20th century. You see, you see my point?

Lex Fridman 1:11:31 I, I see your point and it's a very strong point. In part it's a cynical point which is that the rule of a centralized state is more negative than positive.

Balaji Srinivasan 1:11:50 I think it is a, it is like, it's like nuclear energy, okay? It's, it's or it's like fire. It is something which, um, you're going to keep having it reform because there's, there's good reasons where you have centralization, decentralization, recentralization, but power corrupts, absolute power corrupts absolutely. And you just have to be very suspicious of this kind of centralized power. The more trust you give it, often the less trust it de- it deserves. It's like a weird feedback loop, right? The more trust, the more it can do. The more it can do, the more bad things it will do.

Lex Fridman 1:12:24 So okay, (laughs) there is a lot of downside to the state being able to track you.

Balaji Srinivasan 1:12:31 Right.

Lex Fridman 1:12:32 And history teaches us lessons when at a large scale especially in the 20th century at the largest of scale a state can do, commit a large amount of murder and suffering.

Balaji Srinivasan 1:12:42 And by the way, history isn't over. If you think about what the Chinese are building on this, right? That surveillance state, it's not just tracking your name, it's tracking everything on you, you know? Like WeChat is essentially like it- it is all the convenience and none of the freedom.

Lex Fridman 1:13:00 So that's the downside but don't you... The question is ab- is I, I think probably fundamentally about the human na- uh, nature of an indi- individual.... of how much murder there would be if we can just disappear every time we murder-

Balaji Srinivasan 1:13:16 Well, I mean-

Lex Fridman 1:13:17 ... at the individual level.

Balaji Srinivasan 1:13:18 So, the issue is basically like, once, once one realizes that the moral trade-off has two pulls to it, right? And moreover, that basically centralized organized murder has... I mean, if we add up all the disorganized murder of the 20th century, it's probably significantly less than the organized murder that was, that these states facilitated, right? And probably by, you know, R.J. Rummel has this thing called democide, right? And the thing is, it's so grim, right? Because you know, the saying, like, one, one death is a tragedy, a million is a statistic, right? These are just, like, just incalculable tragedies that we, we can't even, you know, uh, understand. But, um, you know, nevertheless engaging with it, like, you know, I don't know is the ratio 10X, is it 100X? I wouldn't be surprised if it's 100X, right?

Lex Fridman 1:14:02 Yeah, but have you seen the viciousness, the negativity, the division within online communities that have anoni- a- anonymity? People-

Balaji Srinivasan 1:14:11 So, that's the thing, is basically you, there's also a Scylla and Charybdis. I'm not, you know... When you, when you see what centralization can do, and then you, you correct in the direction of decentralization. You can over-correct with decentralization, and you get anarchy, and this is basically then you want to re-centralize, right? And this is the, you know, uh, I think it's The Romance of the Three Kingdoms, uh, "The empire long united must divide. The empire long divided must unite." That's always the way with it, right? So, what's gonna happen is we will state certain verbal principles, right? And then the question is where in state space you are. Are you too centralized? Well, then, okay, you want to de-centralize. And are you too de-centralized? Well, then we want to centralize and maybe track more, right? And people opt into more tracking because they will get something from that tracking, which is a greater societal stability. So- so, it's kind of like saying, "Are we going north or south?" And- and the answer is like, what's our destination? Where's our current position in the, in the civilizational state space?

Lex Fridman 1:15:06 Well, my main question I guess is, um, does, uh, creating a network state escape from the, some of the flaws of human nature?

Balaji Srinivasan 1:15:15 No.

Lex Fridman 1:15:15 The reason you had got Nazi Germany is a large-scale resentment of different explanations for that resentment-

Balaji Srinivasan 1:15:22 Yeah.

Lex Fridman 1:15:22 ... that's ultimately lives in the heart of each individual that made up the entirety of Nazi Germany. And it had a charismatic leader that was able to channel that resentment into action, into actual policies, into actual political and military movements.