Balaji Srinivasan: How AI Will Change Politics, War, and Money

So polytheistic AGI, I think, is one very useful macro frame. Means every culture has their own AGI, and eventually every culture has their own social network and cryptocurrency and AI. You know, the AI is sort of like their oracle that's at the center of society and they've got their deterministic law with cryptocurrency and their probabilistic guidance with AI, the social network that binds the whole thing together. So those three technologies are like social technologies that are almost like the reactor core of the, of the network state of a modern Internet first society.

Slate. Martin and I were talking off, offline about how amazing your thread was on AI. And you know, normally or often, you're a crypto guy, you're a network state guy, but, you know, you're, you're, you're a technologist.

Sure.

And you've been thinking about AI for quite, quite some time. So why don't you trace us through your, your evolution a bit?

Totally. Yeah. So, um, you know, actually Martina and I are roughly contemporaries at Stanford. I got my PhD in 05.06.

I got mine in 07.

Yeah, yeah, I'm, I'm a little gray down over here and Martin's a little gray up over here. And we're complimentary, complimentary grays.

We both have that ambiguous kind of Middle Eastern look.

Yeah, exactly.

That's Middle Eastern.

That's correct. That's exactly right. That's right.

So.

So we're both men of a certain age. I think it's fair. This is, as the phrase goes. And the funny thing, I'm actually, you know, I, I taught machine learning and computational statistics and so on the context of genomics at Stanford for, you know, the mid-2000s and founded a DNA sequencing company. And really, I would say, like my original career expertise for 10 years, that's all I was doing, was in a sense, ML full time. But then I got into crypto right. In the early mid 2010s, just as a deep learning revolution was getting underway with Imagenet and, you know, all that series of papers in the mid-2010s, early 2010s. And so I've suggested my thought process. I have a foundation in probability and stats and multivariable calculus and blah, blah, blah. And so I'm conversing with the space, the thing I will say, and I'm sure if Martin has thought on this, and we'll get to the specific Tweet and the 2000 and tens, you know, we were all tracking diffusion models and Language models and so on, so forth. And it was improving. But, you know, like style transfer, for example, the mid 2010s was working by that time, right. And, you know, GPT2 and so on was interesting. It could, it could kind of blurt out like a sentence. But I admit that I never really thought it was going to get past, like, Markov Chain E like stuff, you know, I was surprised at. At how much better GPT like Dall E was a hint, you know, in. In the early, early 2022. But I was surprised at how coherent ChatGPT was. I think everybody was. But it was like a huge jump up from what it was before in terms of sort of being Markov Cheney. And so I've been kind of observing over the last two years being very originally very deep in machine learning, then very deep in crypto. But you can't be deep in everything. You can't be deep in. Maybe Elon can. Okay, aside from Elon, it's pretty hard to be at the cutting edge of so many different fields at the same time because they're deep fields which have a lot going on. So I, you know, with respect to AI, I think there's like, several realizations I've had for the last few years in terms of the unarticulated limitations of the space. Right. Some of these I think I sort of. I think I came with relatively early in the history of modern AI and others I think I've come to more recently. But let me kind of enumerate them in no particular order. Okay. I made Martin jump at any time. So the first. Is that something that motivated, you know, I think both Eliezer and Altman, a bunch of other folks who built OpenAI, was almost like the same kind of sentiment that motivated people to build, like the Sistine Chapel, which was sort of like an implicit Abrahamic monotheism, which was like summoning God. Right? Like, you know, communing with God, but in the Abrahamic God sense. Right. Of like also the vengeful God who would turn you into paperclips, like turning you into pillars of salt. Right. That was sort of. That was an implicit thing that was behind. Right. So because when they. When they talk about AGI, they talk about AGI as implicitly a unitary thing. We will get to AGI and then it will go to infinity and will be like, you know, raptured singularity kind of thing, even though that's implicit. And, you know, the way I. Because I've been thinking so much about crypto and other kinds of things. I was like, well, there's a different sort of implicit school of thought, which is polytheistic AGI, right? Do we have, rather than the vengeful God, do we have War of the Gods? Do we have many superhuman intelligences, all from different cultural backgrounds that have the mores and the values imprinted on them? And very early on I had, I had a tweet that said, at a minimum, there's going to be American AI and Chinese AI, and if we're lucky, there'll be decentralized open source, you know, crypto style AI. Because at the time the American AI was woke, highly woke. This is 2022. And I knew, I knew that China was going to like stop at nothing to copy it. So I knew we were going to get at least two. And I thought we might get N if we had, if we were lucky enough to get decentralized AI. And that wasn't obvious at that time because the cost training models was so high and OpenAI was so far ahead. It took a while before other people caught up. But now it's very clear that we're going to have lots and lots of high quality, open source, decentralized models that, like, a new one comes out almost every week, you know, and China's going to work hard on this. It's a big thing, like all the deep SEQ models, et cetera. So polytheistic AGI, I think, is one very useful macro frame which takes away some of the sort of AI apocalypse tones, I think, because I don't think the image generators or, you know, text chatbots are going to cause the, you know, destruction that people had. People thought they were going to like bust out and, you know, do systems programming. And Martin can speak about that. Martina and I kind of were talking about that. There are certain limitations. It sound more clear, actually that they're worse at systems programming than they are at visuals. Come back to that. So a polytheistic AGI. And so what does that mean? That means every culture has their own AGI and eventually every culture has their own social network and cryptocurrency and AI, which is, you know, the AI is sort of like their oracle that's at the center of society and they've got their deterministic law with cryptocurrency and their probabilistic guidance with AI and the social network that binds the whole thing together. So those three technologies are like social technologies that are almost like the reactor core of the, of the network state of a modern Internet for society. Right. And it'll be customized for each different kind of group. And certain things will be disallowed and allowed, like image generation might not be allowed in subcultures or nsfw, whatever. All that kind of stuff would be, would be tweaked. And so, okay, that's one concept. The second concept, you know, Eliezer Yudkowski, who, by the way, you know, even if I disagree with a lot of his stuff, did do a lot to promote AI and people getting into it and so, and so forth. So even if I disagree with bombing the data centers and various other kinds of things, I give him partial, significant, partial credit for getting people motivated to look into the space. He was definitely into it in that sense. So, like, directionally there, there was something, you know, I just try to see the right side of things. But a second big concept that I think, I really disagree with Eliezer on, that I think is being borne out is the idea that AI could just cogitate for millions of years and figure things out and it could, could outmaneuver you all the time. And we know that's not true because turbulence, chaos, cryptographic equations are not like that, right? You can come up with turbulent systems, chaotic systems, where you simply, with finite precision arithmetic cannot forecast out indefinitely. In fact, you get fracturing and breaking and cryptographic hashes are set up in such a way as, well, to be hypersensitive to initial conditions where a small change of one character can get a totally different MD5 sum or something like that. And so those are situations. And if you wanted to, you come with a thought experiment where you inserted turbulence or chaos into your decision process. Sort of like, you know, you shake a turbulent clock or something like that before throwing a pitch. And the AI wouldn't be able to predict your actions, right? And that's like just a simple experiment in real life, just, you know, the flow of a fluid is turbulent or, you know, so. So that actually put bounds on what AI can predict, right? Quantitative physical and mathematical bounds on what an AI can predict.

Can I just add just a little bit of color here? Because I think this is cheap. I think we need to call out why. So the way that you describe AI as gods and monotheistic and polytheistic.

Sure, sure, sure.

It's great at describing how human beings view AI, right? But in reality, we're talking about software running on computers that are bound by those limitations, right? So I don't view them as gods. Personally, I'm a. You know, I got my PhD in Systems. I view them as like System software. And I, I actually think the original sin in all of this AI, you know, anthropomorphic fallacy started with Bostrom. Right. It was, it was one of these kind of thought experiments where, you know, when Nick Bostrom wrote Super intelligence was what, 2014. And you know, listen, he was talking about this platonic ideal of AI and this Platonic ideal of AI just happened to be able to recursively self improve and just happened to like have these kind of like super physical things that like no AI today has. But it happened that the conversation around AI started then. And then it just coincidentally we had these LLM show up four years later. And so somehow our thought process on these two things dovetailed. Right. And so people would take all of these kind of mental ruminations, these thought experiments, and they would apply it to like actual systems. And I think the problem with doing that, the problem with taking some Platonic idea, whether it's Bostroms or it's whatever Abraham, the Abrahamic view of God or any kind of religious view, is that it is very blinker to the limitations. And you've pointed out a great limitation. We've been doing computer simulation forever. We totally know the limits of simulating physical phenomenon, particularly chaotic systems. We have actually very hard bounds on these, by the way. It's not just the limits on the size of a computer word or an integer or something like that. They're actually very strong limits on time and the amount of compute necessary, et cetera. Right. And so we know these. And so you can talk about these in two ways. I think for this conversation we should do this like Balaji, you're so good at talking about like, what is this platonic ideal and how should we have a mental kind of model for this for like the non computer specialist? And what I would love to do as we go through this conversation is talk about like actually these are still bound by computer systems. We know the limitations of computer systems. And so let's see how those bound it. And you just beautifully did both of those things in the same one. I just want to make sure that we tease apart both of those things as we have this conversation.

Totally, totally. I feel I can kind of speak both languages here where I underst, I understand where those guys are coming from. And because I kind of am like also a tech radical, you know what I mean? Right. And. But I'm also a tech pragmatist. So I like almost, I think I straddle that boundary.

The danger is if we don't say that this is a Platonic ideal. People will map it to existing systems incorrectly. And that's exactly what happened in 2020. 2021, which is they. They took this thought experiment that started with Bostrom, and it was a. You go back to the original book, you're like, listen, this has nothing to do with real systems. And they applied it to a real system. And I think that was kind of the. Originally.

Totally. I know, I know. Well, but. But let me poke on that a little bit. And then both. Both poke and then also defend your point. The Turing Test was a thought experiment that was a Platonic ideal. No. No reference to neural networks, no reference to implementation details. And yet it served as something that went from a thought experiment to an applied thing with, you know, the Levchin Gaussbeck test, the Captcha was like, v0 of that, you can argue, and became commercially important. And now, obviously, AI's blown past the Turing Test. It can, you know. Right. Beat people in this. And then, you know, the Chinese room, John Searle's thing. Right. About machine translation. Another. Another thing that was like a Platonic ideal. And there's other things that are like that. Like six degrees of separation, which social networks actually made real. Right. So. So I'm not necessarily against Platonic ideal.

Thought experiments are very, very important. I just think that we need to be very clear when we're having a conversation not to conflate them with, like, the actual system.

Yes.

And I think. But that's the thing that we've kind of fallen into in this conversation. Not you and I. I just think the broader discourse inside of us has.

Fallen into the broader conversation. That's right. So. So now your point, which is a very good one, is these are real systems, and they actually have real limitations. I think one of the more interesting things for me over the last few, you know, years, months and years, has been defining exactly where those limitations are, because I think where they landed was kind of counterintuitive. Right. So let me give some of them that. That I think about one of them is that you have this decentralized AI rather than AGI. That alone, I think, kind of nukes a bunch of the concepts of we'll get to AGI and just win, because it's kind of clear that there's just like a rapid onrush of new models, and it's like more of a continuous kind of thing. The fast takeoff scenario didn't happen. Right. On the other hand, I think that it's not like, can an AI write a sonnet? It can Right. It can do it better than most humans. Can it write a screenplay? It can do that again better than most humans. There's a lot of things that we thought as maybe harder. We thought maybe like locomotion or something would be easier to solve than what we think of as higher cognitive functions. But it's actually the locomotion that's still harder in some ways. Right.

I think that one now in retrospect is easy to explain via economic equilibrium. We just didn't have the right model.

Why do you say that?

Well, so for example, when you're competing with a human brain on 3D navigation in space, you're competing with whatever the 4 million year old mammalian brain and a body that's been running away from predators and picking berries for 4 million years. It's incredibly highly evolved. When you're competing with a prefrontal cortex, which is like, you know, the language learning and creativity.

How old is it?

250,000 years. And so if you take this question from an economic equilibrium standpoint, you're saying, well listen, do you want to like compete with like the most evolved system that's solving the much more difficult problem, which is much higher dimensionality, has all of the chaotic has to deal with like, you know, chaotic nonlinear systems like you said, or do you want to deal with the very new evolution that deals with kind of a much denser space that actually does a pretty good job with linear interpolation. The problem set actually works very well with linear interpolation, et cetera. And so I think you're totally right, it was not obvious. I mean, listen, AI has all been solving what we thought was the easier problem. But our fallacy is easier for humans because we're really good at it, because we've been doing it for a really tomorrow. And it turns out the harder problems are just harder for us because we've only been doing them more recently. So this is a great case of where it did not like, like our intuition on the problems to solve with the wrong ones because of, of our own kind of anthropomorphic fallacy. Right. Our own notions of what problems are easy and hard.

Yeah, I mean, okay, true, I would say, you know, one of the, one of the surprises I had, that's why I was interested when you said economics. One of the surprises I had with GPT3 and ChatGPT was how far you could get with language. Yeah, like before the CHAT GPT moment, it wasn't obvious to me that just then after you're like, okay, language is sophisticated enough to encode almost any concept about the world.

Right. Or rather, if I put you in a very dark room that's of arbitrary construction and I describe how you navigate and pick something up, I think you'd have a very tough time doing it.

No, no, I know. What I'm saying though is basically, if you generalize language to be streams of symbols, you could send telemetry to somebody, like, basically. I think I was surprised by how many concepts were encoded in language that could be learned. Even to the point of rough world models of the map of the earth or the proximity of things. You can back out those kinds of things.

The distinction is it could have been. I totally get the distinction is it could have been the human being, the human mind that looked at the world, did the reasoning and created the world model. And that is cast in language. And then language is.

Yeah, the same thing, but from the.

World to the world model was the human and then everything else and then the language. I agree.

Yeah, that's right. But it was a little surprising to me that you could get that far with language models as opposed to like spatial reasoning, as opposed to, you know, like that predict next token would get as far as it did. That was surprising to me that that was the angle would not have been. The reason is you and I both did so much stuff on Markov chains and conditional random fields and all that kind of stuff. And of course the transfer is a different architecture. But I just wouldn't have. I just wouldn't have believed that that method taken to. I mean, another thing I think was very counterintuitive for me in the late 2010s was double descent, because as a classical machine learning guy, like, that's just very counterintuitive that you could go past over training back into like, you know, good regime.

But I want to.

I want to. So I want to actually talk about one thing you did say, which is self replication, Right. I don't think of that as a forever constraint on AI. I think of that as a today constraint. And the reason I think of that as a today constraint is they're not embodied. So they're not scripting robots because they're not scripting robots. They can't build data centers and minds and replicate themselves and so on and so forth.

Right.

And you know, the whole concept of consciousness initially was that consciousness evolved so that like, you know, you're running away from like a boar or something like that, and you have a model of yourself and then there's like a branch ahead. And if you Have a model of yourself, you know, whether you'll fit under that branch or not. Whereas if you just had a generic model of like the more self conscious you have, you can sort of simulate your run under that branch whether you're going to die or survive. Right. And so that's like one theory on why consciousness arose. To help your survival replication.

Stand outside of yourself to be able.

To see the experience. That's right. So right now AI does not really have goal setting, it doesn't have reproduction, it doesn't have embodiment, and it can't act independently of humans. This is one of the big things I think that people were really scared about in late 2022 and they've calmed down on this that you and I both poked on. Martin, I think was, is this thing going to jump out of the box and you know, like code itself? And the thing about it, the reason that that was, now we laugh at that. But the reason that I think that hasn't happened is AI can't prompt itself yet. And prompting, I argue, is actually a much harder thing than people realize. Because did you see, did we talk about my analogy of like the spaceship?

Talk about it. That's definitely worth going through.

Okay, so let's say you have a really fast spaceship, like close to speed of light or something. You still have to point at the phi and psi coordinate on the surface of a sphere, you know, in coordinate space as to where you're going to point that ship. And if you're going to take it on a, on a journey, then you've got waypoints like here's this heading and here's that heading, and so on and so forth, like a series of phi psi pairs on the surface of sphere. Okay, so that's only two floating point variables, you know. Right. By contrast, if you take, I mean, how high dimensional is the vector that you're giving as input when you talk about a prompt? Right. If you just take, you know, UTF 8 code points, I mean, or just even ASCII, right. And you have like a few words that's, I mean, that's much higher dimensional than just a, like a vector of, of two floating point variables. Right. So you can get to a very high level of dimensionality in terms of the direction vector you're pointing this AI spaceship in. Right. And so a prompt is a very high dimensional direction vector. Even if you account for the fact that many potential prompts of just strings of random characters wouldn't, wouldn't be interesting, it's still a very, very High dimensional vector. So it's like you've got a fast spaceship, but you still have to point it in a direction to go somewhere. I think that's a good analogy.

Well, I think there's one more level of complexity you have to add to your analogy, which talks about how difficult it is to make these things kind of, let's call it autonomous, which is closing the control loop. So let me try, let me, let me, let me try and build on your analogy, which is.

That's the verifying part. Go ahead.

Yes, so it turns out that the directions that you point in, it has to understand, right? Like that has to be in distribution. So you can't point into a direction it doesn't understand because it does the worst thing. If you point in a direction it.

Doesn'T understand, it just crashes right into that wall or whatever. It just tries to do it.

It's a random direction. It's the worst thing ever. Right. And so the problem is if it's producing a direction that has to go into itself, it doesn't know what it knows and it doesn't know what it doesn't know. Does that make sense?

So it could produce, in fact, it's optimized to fake it.

Yes. And so if you could tell it, if you could say, hey, listen, produce a bunch of directions by feeding the last direction in, you have no idea of a direction it spits out is going to be in distribution when it comes back in. And that's what closing the control loop means. And it's a very tough problem. I think this is such an important point for us to go into because in theory you can close a control loop on these things. But as scientists, we want bounds on what that means. Right. For example, clearly you want to gather new information to update your model. Then we want bounds on how much information you need to gather. It turns out that model was trained on everything humans have ever gathered. So it is like the incremental experiment going to update that maybe information theoretically says probably not, but we don't even have bounds on these things. Now you said previously, well, when it comes to a chaotic system, we know that computers, you know, it takes a long time for them to compute a nonlinear system.

Actually, you know, I just want to pause you there. You just gave me an idea for a great prompt, which is what areas do you feel your knowledge is the most thin on?

This is the key. Like does, does, does a model. Does a model.

That's a great point.

Yes. Does a model know to what extent is in distribution or is out of distribution.

Yeah, I'm going to try that one actually, by the way.

This is, but this is the key. This is the key. Self reflection is the key. Because if it produces an output that's out of distribution, then of course you have error and then you're not there to kind of nudge it back.

Yeah. So it's like real time events, obscure or niche academic fields and specialized subfields behind paywalls, local and regional information, human emotion, intent or experience. Private or proprietary systems. Actually pretty interesting, quick off the cuff response here. Right. So you can, you should ask if.

It can always produce a response to which it has a lot of data.

Always produce response. What do you mean by that?

So the question is, is if you're, if you're closing the control loop, it's going to spit something out that you're going to feed back in. Right. That's the whole point. So the question is, will it always spit stuff out that when, if you feed it back in, won't give you nonsense.

Oh, I see, yeah. I mean, well, people have actually tried like that, that experiment of take the image and just exactly replay the previous image and then it like morphs into something totally different. So another kind of angle I have on AI is prompts are tiny programs, which is more common today, but I think at least I articulated it went viral a while ago and. But they're programs in a hidden API because normally you have an API that is fully documented but very error intolerant. Right. Prompting is the opposite. It's completely undocumented, but it's highly error tolerant. It will usually do what you mean, but the better your vocabulary, the better you could prompt it. So now art history is an applied subject, right. Political, like knowing a vocabulary of like Cezanne versus Picasso and so and so forth, you can actually pull up the style that you want on demand. So the broader your vocabulary, the broader your subject knowledge, the more you can get out of it. We're in the age of the phrase which is the prompt, the 140 character tweet and the 12 words for your crypto password. Right. These, these phrases of power in AI, in social media and in crypto, just unlock everything. So the better your vocabulary, the more you can do.

Right.

So I think of prompts as tiny programs and actually one of the things I've gotten in the habit of doing is writing. It's total opposite of search. You know, with search you, you learn to type things in keyword ease and you sort of figure out the, the word that has the most specificity. TF idf, you know, on the page or whatever. I will write sometimes these long memos to an AI and then continuing. I know maybe you don't love this analogy, but I think it's funny. Continuing the polytheistic analogy. I'll give them to Brahma, Vishnu and Shiva. Okay. So I'll give them to Chatgpt and Claude and now, you know, Grok and what have you. Right. And I'll consult all the gods and then I'll make my decision on that basis. Right. And then sometimes have them argue with each other. Right. And why do I say kind of half jokingly gods? Because, you know, like the Hindu kind of frame on that is not like the, you know, fearing God thing. It's not the same kind of thing. But in a sense it is a superhuman intelligence that knows everything about your culture. And if you ask it the right question, it can tell you something that you didn't know. But in like the Hindu tradition, they're not infallible. Right. Like that is to say it's not the same as the all knowing, all seeing. It's like this interesting. It's like more like superhumans, almost more like superheroes. People will argue with me on about that, but I think that's more true.

Right? Yeah.

It's like actually the Norse tradition is similar to the Hindu tradition in some ways. Right. Where you have, you know, the gods were not infallible, but they were superhuman.

I think this is, this is a great. And it's a useful framing. Just remember that when it comes to computer systems, we can put formal bounds on them. We can do this information, theoretically we can do this computationally and like that's, that's going to come. And once that happens, we will understand these systems fully and it'll be very hard to think of them as gods at that point.

Well, that's right. I mean, that's the thing is actually what's interesting is that the, the interpretability work that Anthropic and others have done and the work on like grokking or what have you. Right. That's actually really good stuff because the, you can pick apart neurons, you can pick, you can find the golden gate neuron if you, you know, if you found, if you saw that kind of thing, right. You can dial that up, dial that down. You can start actually taking apart these AI brains in a way that hasn't happened before. A few other kinds of things. Right. So like the thread that you guys, that, that thread actually summarized several of my recent. I Should probably put this into a post so it's like kind of there for the record. So I'm just going to do a bunch of these then maybe get your thoughts. Right. Okay, so first concept, and in no, no particular order, AI doesn't do it end to end, it does it middle to middle. So the business spend. So basically you have to still prompt it and then you have to verify it. And people talk about prompting, but they talk less about verifying. And Karpathi and I had this good conversation a few weeks ago where basically AI is going to create massive numbers of jobs in proctoring and verification because it's so good at faking things. So one of my other kind of concepts is AI makes everything fake and crypto makes it real again because AI is a probabilistic technology and crypto is a deterministic technology. And so like crypto is in some sense what AI can't fake. It's like the hard cryptographic equations. It can't fake a bitcoin private key, it can't fake even an on chain nft. That's what AI cannot fake. And so that's, you know, like, like the hard barriers. Right?

I mean, I generally agree, I don't, I mean, I don't think crypto solves the grounding problem. Right? I mean, it's a, it's a, it's a mechanism you could use, but it's a mechanism like the grounding problem is.

Well, so you mean by grounding problem is grounding in reality?

Yeah, yeah, yeah. The actual physical grounding problem. Ingest problem. Yeah.

All right, I, so I disagree with you on that and here's why. Or let me give you a counterargument at least right now. Let's say, just give a concrete example. Let's say you asked Perplexity to summarize the ftx hack in 2022 when it do so, among the citations it would give you would be links to a block explorer, right. That would actually have on chain data that you can cryptographically verify that this transfer of these funds happened at this time. And if you want to go even further, you can actually pull out the digital signatures and the hashes and the timestamps from that block explorer. Right. Okay, now here's my argument. My argument is that works for financial data, but what's happening now with Farcaster and other kinds of things is with increase in block space, you could put more and more kinds of data on chain and we're going to have to, because you're going to need Crypto instruments. And you're going to need cryptographically hashed posts and crypto IDs to know that it was posted by a human or to know the data wasn't tampered with. So more and more kinds of data are going to go on chain and then that will eventually mean that an AI citations are to on chain data, which is both financial data and social data. And so then at least it will map back in terms of grounding to an on chain cryptographically provable assertion of some kind. And you might say, well at least that'll be an assertion at the metadata level. Like we can prove that this digital signature made this assertion at this timestamp with this probability.

I'm talking about real world grounding. Like I say something, I am a human being, you know, you have no idea what I said is true or not true. You know, there's a geographic place where there's a picture of the geographic place taken from a 1970s photo. Was that doctored or not? I mean like, like the actual physical world grounding. Just because you can't encode, you know, digital data yet at this point for the physical world now and over, it's a great mechanism to do that once we can solve the ingest problem.

But this just so, so let me, let me talk about something which is happening now that I've been kind of funding on the side. It's not a full solution, but it's I think a partial solution which is. So crypto instruments, the idea would be that when you capture like a frame of data, right? For example, sequencing machines like DNA sequencing, when the data is coming off the machine, it's like TIFF files that are actually image data that gets processed into and many other kinds of instruments basically have a stream of data coming off the machine as you're capturing it. Cameras are like that, right? Okay. So you could, and there are things that do this already, take a hash of that and post it on chain at that time and do that.

Right.

And what that would at least say is that that frame of data existed at that time. And so if you had something that was like a scientific experiment, right, you know, like a, like a pre registered double blind trial or something like that, you could have not just a crypto instrument, you could also have other people with proof of humans there who have a sort of attestation ceremony. And now you have a number of different kinds of on chain data that start to get harder to fake in a corded way. Not impossible totally.

As soon as you get it into the system, then crypto is a great mechanism for ensuring kind of end to end guarantees. It's just the data ingest problem is maybe it's a long standing problem in computer science. That's right. And over time everything you're saying is going to be more and more true because over time we're going to be more and more incented to make sure the stuff going into the system is true. And so I just. Totally agree. Listen, I'm an old school networking guy. For us there's the Internet and there's the stuff that goes in the Internet and then you just kind of use different mechanisms for both and it's worth calling out. That's all.

Totally. That's right. Okay, great. So, okay, so next kind of concept maybe to discuss. And I think this is a useful like division that this was a relatively recent point that I made to myself that thought was useful. AI is good for the visual and less good for the verbal. What do I mean by that? So when it's generating images, when sharing video, when it's generating user interfaces like Vercel's V0 or Replit's user interfaces, the great thing about them is you can instantly see them. And with the GPUs that we have in hardware, you can verify cheaply whether they're good enough.

Right.

Because you can just instantly get the gestalt of it. Right. Whereas when it's backend code, when it's legalese, when it's like, you know, mathematical equations, you have to slow down and use system two thinking, not system one. Right. And it's not just your gestalt impression. You have to actually go line by line and check whether it's right. And that is actually the expensive step, the verifying right over there. So that's actually, I think that's a non obvious thing where the more front end and video and visual it is what you're doing, the easier it is AI. And now the interesting concept is how much of that. Go ahead.

One more thing which is like for me again like I spend most of my time in like software and engineering, the big distinction is stateless versus stateful. Right. So if you're generating code that's going to have semantics that evolve while you're running it, it's just impossible to spot check. Like some things are computationally irreducible. You actually have to run the computation to know the answer. So I told like the image is the perfect example. It's visual and it's basically stateless. Like all of the State is there. There is no kind of runtime semantic. Totally agree.

That's right. And whereas even a relatively small snippet of backend code could have a fairly complex finite state machine essentially underlying it, or even infinite state machine, you know, like. Yeah, it could be, it could be very. That's right. And so simulating the time dynamics of that, you have to use something different, maybe formal verification if it's algebra, you.

Know, or you'd actually have to run it if it's computationally irreducible. Like there's no way, there's no way to statically do it. I mean, and so it does reduce to almost like the computation verification problem, which is this long standing problem in computer science forever right now.

Now formal, formal verification, at least for a subset of programs has become commercially viable for smart contracts because they're so high value and they're so small that they actually are. It's worth doing that on. Right? And yeah, it's not going to be, it's not going to work for the general case, but you can do a constrained case. But okay, so that was one major division, visual versus verbal. Another when you're getting to like stateful and so on, is I think there's certain. The limits of AI are the things I'm interested in where you draw like a fine distinction of what it can do are very crisp. And what I think AI is particularly bad at that people are trying to make, trying to use it for, that they're going to fail on, in my view, is when they try to use it for markets or politics. And let me explain why I say that for systems that are time invariant, like the mapping of an image to the label cat or the rules of a game like chess or checkers or even go, or you know, something where there's like a, like a, sort of like a static rule set or static mapping, right? Then you can do the train test paradigm and train a model and so it's so forth. However, when you have something which is time varying, especially rule varying and adversarial, like markets are or like politics are, then the same trade will eventually quickly start resulting in a loss. And by the way, the other guys are also using an AI on you, right? And so this decentralized AI again, right? And so what that does is actually argues that the CEO or the infl or the creator who is constantly sensing the market or sensing the political wins and has a thesis on it based on human nature or other things or what have you, actually is the sensor that then prompts AI. And that's a job that is hard for the AI to do at a really deep level because it's time varying, rule varying, adversarial.

And it goes back to what you said at the very beginning, which is if you look at these type of equilibrium, they're complex differential equations which are nonlinear. And so in order to predict what's happening, we'd have to do this nonlinear extrapolation which we know that these things are not very good at. Right. That's what you use.

In fact, actually what's interesting is I wasn't even thinking of the stock market as complex, different.

But you're right.

Even in the subset of time wrote this great.

On the fact that these things are these chaotic. They're super chaotic.

Yeah, you're absolutely right. That actually it would be a useful thing to show just with a toy example a chaotic system that is time varying or another one's adversarial dynamic. Now the thing about that though is to argue against my point, they've gotten AIs that are actually pretty good at StarCraft. Right. Which are. It starts to stretch the boundaries of what I was saying because it's definitely adversarial. Right. And it's like more time varying than chess. You know, it's, you could argue it's not rule varying, but it's, it's, it's time varying. But so. Okay, let me, let me go to another point here. Right. So another concept is, I think so the commercial implication of that, that point on prompting and verifying is business spend moves towards prompting, proctoring, verifying, basically checking all the stuff that AI can generate. That's going to be a huge, huge, huge thing. And that, that maps to kyc. That maps to like in a bad way. You know, the glass cases in Walmart. Right. In a sense, a low trust society is spending more and more and more on verification and proctoring and so on. Right. Okay, next, AI means amplified intelligence, not agentic intelligence. Because the smarter you are, the smarter the AI is. Better writers are better prompters. What are your thoughts on that?

Yeah, I mean I think it's interesting in the coding space that we actually start to have numbers on this now. Like if you actually look.

Oh, interesting.

So if you actually look at like relative productivity gains, it just turns out if you're a more senior developer, you will have better productivity gains. And the ar. I mean.

Oh, I hadn't seen that graph. So you actually. It is something that makes the smart.

Smarter basically, but also on a relative basis, which is really surprising. Right. But if you think about it, it's actually not surprising. It's like, you know what the fundamental trade offs are, you know what to ask, you know how to interpret the results, you know how to throw away bad stuff when it's bad. And so clearly if you kind of know what you're doing, you can both verify to your point the output, but you can also be more specific of your ask. I think it's really important for all of us to realize that formal languages came out of natural languages, not the other way. Right. Like if you could explain all of this stuff in English to each other, we would, but it's just really inefficient. So we came up with like more efficient ways to explain trade offs.

Constrained.

More constrained.

Yeah, basically constrained languages that reduce to ambiguity.

It's literally strictly an efficiency thing. Right. And so like someone that knows how to speak these formal languages to the, to, to the models you're going to get is going to articulate what they want better and just going to be able to interpret the results better if the response is formal. And so it is kind of a nice codification of exactly what you're saying.

Yeah, I mean, the thing about it is AI means everyone's a CEO because you kind of speak to the AI like you do to a, like to a great AI, you speak to them like in some ways to a great employee where you give clear written instructions and then you can verify the output and then, you know, like, so it actually kind of turns management into a skill that it hyper deflates the cost of trying one's hand as a CEO or as a manager, because you have to give those. And so the better you are in communicating what it should do, often the more people you can manage and so on and so forth. By the way, this gets to kind of the next point, which is AI doesn't really take your job, it takes the job of the previous AI. Okay. And what I mean by that is you now have a slot on your roster at every company for an AI image editor, an AI text, you know, chatbot thing, an AI code, you know, IDE thing, and so on and so forth. And each new release of GROK or Claude or whatever competes against chatgpt and GROK and Claude. Right. And so the AI takes the job of the previous AI because they're complementing. You kind of have a whole raft of AI augmenters that are augmenting your humans, but those AIs are competing in AI space to a large extent with the previous AI, because once you've onboarded an image generator into your flow, then it just keeps improving and you start using it in more places. But it's an AI taking the job. The previous AI. Let me know your thoughts.

Can I actually. This is an adjacent to what you're just saying, but can I actually push on something you said previously because I actually agree with your polytheistic view of the world. I totally agree. But let me just provide the counter argument for us to niddle on in this opinion, which is have you seen this kind of thing that all the AIs that you ask to produce a random number produce the same number. Have you seen this like four?

Yes. It's like 46, seven or something like that, right?

Yeah. So one thing that to me was non intuitive but remarkable about these models is how easy they are to distill, which is as soon as someone creates a leader, everybody uses that leader and kind of sucks the life out of it. And then all the models kind of converge on it very quickly. Which you could argue that this is a counter to the polytheistic.

Yeah, kernel intelligence is basically the. Yeah, there's, there's like a core, maybe.

There are ten, maybe there are a hundred AI, but it just turns out they all have the same capability. So like is it, is it, is it just like a technicality that they're actually different and they've all learned from each other?

Well, so it's an interesting question and my view is, and I'm not called this a strong view yet, but my view is that's almost like the human body plan and spinal column and then you differentiate on top of that core spinal column maybe, you know, and so this like it's like you'd have some sort of. Just like every human to first order can see and speak and hear and so and so forth. There's.

And.

But then some people have, you know, much better vision or they have much better speech or something like that. Right. So there may be some distilled kind of thing. Oh by the way, another interesting part on onto what you're saying. In general, text on the Internet is not emitted equally by every group. For example, liberals tend to write more text on the Internet and conservatives tend to be more visual than verbal. So is actually, you know, you'll have an ideological skew that's actually hard to unskew because it's not a. Like the people who are training AI disproportionately are like redditors.

I totally agree with what you're saying. I'm just saying the counterargument, by the way, I agree with what you're saying, but the counterargument would be you just ask the AI model. Conservatives don't write so much, so give me the answer. Using the mediums that they use, or I mean, just all of that information would be in the model. Here's how I would distill kind of my view on this, which is very much in line with what you're saying, which is I think the universe is very complex and I don't think it gives up its secrets easily at all. And I think the universe is full of fundamental trade offs. Like you can't have both, you have to choose A or B. And so these models will align with those fundamental trade offs, right? And maybe it's performance, maybe it's correctness, maybe whatever it ends up being. And so as soon as you want a specific solution for a given problem where it hits one of those trade offs, you're just going to need a different model because otherwise you just can't end up having both. And again, because I know the coding space the best, we see this a lot, right, which is a model that's very, very good for certain parts of code is just not going to be generally good at other things because those are the trade offs made when training it. And I think that this is kind of the future plurality of model.

Is that true? I thought somebody said something. I may be wrong about this, but I saw some counterintuitive result that said that making the AI specialized in one area makes it worse in other areas. Do you see something like that?

Well, yeah, I think so. Well, this is a very big debate. But the debate goes as follows. The first wave of AI was pre training where everything you threw into just got smarter. And so that's kind of a 10 for 10 technical win, right? Just because it'll be as good as writing code, it is as sonnets. But as soon as you're doing RL where you're training it in a specific domain with a specific verifier, you are likely losing other areas. So you make it really, really good at playing chess, it's going to be less good at something else like that, like writing piano scores or something. I think the current debate now, and the current data seems to suggest that RL doesn't generalize in the same way. And so now we are in this case where you would have a plurality of models because you're always kind of robbing Peter to pay Paul when you make it good at a certain domain that's right.

Yeah. I think also you know, somewhat related to that like in terms of what domains it's good at and so on and so forth. At least right now I think I'm not sure if you agree with this. AI doesn't really take your job. It allows you to do any job because you can get to like an okay level as like a user interface designer or sound effects or something like that. But you need a specialist for polish and that. You know though I wonder maybe with enough RLHF from specialists maybe that won't be as necessary.

So here's my current mental model. There's two Personas. There's Persona number one is the expert in the space and there's Persona number two is the non expert in the space. So if the non expert in the space is using AI, it's taking the place of the expert, right. And so like maybe you'll ask it and it'll give you some. Let's say I want, I want a 3D asset for a video game and I'm a programmer, then I'm going to ask it for like a nice 3D asset and then it'll give me one. Right. So I'm the non expert. The expert user to our previous will actually know how to access ask it better and likely get better results because it's an actually domain expert and that expert is using it. So I, I think we see both actually if you look in the market you see both of these uses and I think both of them will persist which is if I'm a programmer I.

It's like a, it's like a doctor talking to their doctor and they can just instantly go to specialist language. And so you're right, even if they're specialist RLHF, you may not be able to.

Yeah, that's right. Like why would I actually, why would I, why would, why would I want to learn the entire domain and make all of the trade offs of, you know, 3D design? That somebody else could have done all of that work for me and they could talk to the model in the specialist way when I can just talk to my model using code. Right. In order to very efficiently use a specialist model I would have to become a specialist would be the argument to our previous one. So I think, I think listen, for casual use I can use these models for whatever I want but like to really use them very well. Again to our previous conversation, I'd have to become a specialist and somebody else will have maybe already invested all of that time by the way. If you actually look at products, these products actually have both these distinctions. Some products are very clearly for the casual user trying to replace like, think about like cursor versus lovable. Right? So lovable is like, I'm a casual user, I'm going to create a website, I don't even have to know about code. And that's great. You can create amazing things. Cursor is. I am a professional software developer. I have an IDE and I own ide. Now over time, maybe these things converge. That could be the case. But thus far, these are kind of very different user bases. Right? There's professional coding versus basically casual coding.

I mean, you know, part of it is, which is interesting and a little counterintuitive. It's like if you think about what a computer can do, it can do the job of an accountant, it can do the job of a physicist, it can do the job, you know, but then you clad it in something and it's adding up numbers in Excel and you clad in something else and it's doing simulations for matlab. Right? And so we already know that when it came to logical system 2 thinking that computers are actually really good at that, and now you have something similar where these, these models actually very versatile, but you clad it in the power user interface and you clad it in the, the, the casual interface. And you know, it's kind of implicit contextual prompting as well, probably as well as a system prompt that makes it do those things. I mean, the thing that's interesting to me about something like chain of thought is that and, and I think this is where people were freaking out in late 2022, and maybe they'll still be right to freak out is computers had historically always been good at the logical style, much better superhuman at that. Now they're also superhuman in a sense at the probabilistic style at least of text generation and so on and so forth. And so it's not inconceivable that someone could figure out a way to merge those two, you know, like, like a quantum gravity, you know, kind of theory of, of things, right. Where you take the probabilistic and the deterministic and pull them together.

Yeah, this is, this is a very old school, you know, systems. Part of me thinks that there's a fundamental trade off here which is like you can trade off.

What's that?

Well, I, I feel like you can trade off, you can build a system for determinism and you can, you know, build a system which basically cuts a bunch of corners, but you can't build a system that does both. Right, like if, if.

What about the tool use stuff?

What?

I mean, like AI has gotten pretty good at figuring out when it wants to generate an image, when it's supposed to search, when it's supposed to read a PDF.

Right, so. But now you're acknowledging exactly what I'm saying, which is some things you want the fuzzy thing and some things you want a traditional system, like the tool use.

Basically. Yes, but if you have a fuzzy system on top.

Yeah, yeah, for sure. But then maybe it just becomes a consumption layer and all the hard work is still being done by traditional systems. Right, so the tool, the tools.

But if the fuzzy. The argument, if the fuzzy is. Go ahead.

Well, I'm just saying the full argument. The full argument is if you have a model that does everything, you wouldn't need tools. Right, because. Because tools are literally like the API to the traditional system. And so that's almost like a, that's almost like a capitulation that like some things you kind of want, you know, you know, traditional software to do.

No, I know, I know, but what I'm saying is a hybrid system, if you're just a pragmatist and you don't care.

Totally.

Could a hybrid system actually get there? I can't say it couldn't. Right. And you know, maybe it's something where what we actually need is something like Elon's, you know, billion miles of Tesla driving if we have enough context, not from LLMs, but from pointer movements and mouse clicks on, you know, iOS or other Mac OS. Yeah, yeah, an AIOS could. Yeah.

So your question is, can I have one trained neural net, one trained LLM that can do both? Like the fuzzy stuff and the hypersensitive, like, is that possible? My gut, again, this is total intuition, is that the universe is way too heavy tailed, it's way too non linear and so the state space is too high for that to actually encode all of that. And basically.

But humans can do it.

No, we don't. We use calculators and we use software. Like the whole reason we built the whole.

Yeah, I know, but at the meta.

Meta is because humans can't do it. I mean this is.

Sure, sure, sure, sure. But at some level.

Well, maybe we're just talking about two different things. So clearly AI with traditional software can do great stuff. To me the question is, is, can you have one AI that does all of those things as, you know, without the traditional software, can you build one LLM? And I think the answer to me is obviously. No, but, you know, I've heard arguments that it can.

Yeah, you know, what's interesting to me is so, you know, like, that's why I'd say I'm. I'm a little bit of a. I mean, a little bit. I'm on the borderline of the tech pragmatist and the tech radical right, where I think I always want to try to identify the limitations of the systems today and then see how you could push beyond it. So, you know, like, I consider calculators, which were made by humans, to be ultimately humans doing it. You know what I mean? Like, in the sense of it's like a tool that we came up with. I mean, to give you an example, like earlier I was like, you know, AI is good at visual, but not a verbal. But for example, you can turn audio into spectrograms, which you can then look at visually and at least you can see for radical deviations and so on. Maybe not the entire sound, but you can see. And so can we do Fourier transform, like things on other outputs of AI where we can quickly inspect them visually? You know, is there some grid or visualization kind of thing where we can turn into a visual problem?

So what I would love to.

And so in the same way, yeah, I love this.

So what I, like a thought experiment I have is can I literally create a bunch of audio outputs so that I can like listen to my AI Like I listen to a car because like, you know what, like, you know, when your car.

Yeah, yeah, you can tell if it's rumbling.

There's like, I don't know what's wrong with my car, but I know it's not normal type thing. And I do think we're very good at, very good at atmospheric inputs. And I think this is a great idea. Can we start exposing the internals so that we understand when it's working well and when it's not working well, you know, And I do feel like, for example, we could push on here that, you know, we're very early innings of.

Yeah, like, like colored text, for example, in terms of its level of confidence. You know, stuff like that. Like yellow, red, green. Right. There's a lot of AI UX that one can do. Let me make a few other points that I think are interesting. Killer AI is already here and it's called drones. And every country is pursuing it. So we don't have to care really about the image generators and chatbots. All the worry about super persuaders or whatever is all pretty stupid.

Strong, agree.

Strong agree.

Right.

And the thing is when I push people on this, what's interesting is some of the people who were oh my God, we need to regulate everything, are now actually on the side of we need to build it before China. But the thing is in both senses, first it was safety, then it was security. But it both comes to control. And you might argue that the security argument is a better argument. I think it's a more realistic argument in some ways. But, but the concept that killer I is already here.

Yeah.

Is interesting because they, they put so much stock in the like, oh, it's going to persuade everybody to do things. It's a super persuader. But persuading is statistical and you know, drones are deterministic or at least you know that the guns on a drone or whatever.

Yeah, I mean I think that the interesting question around AI and like attack and defense is does it change the equilibrium? Because the Internet did. So the Internet actually introduced the notion of asymmetry, which is the more that you rely on it, the more vulnerable you are. So to wit, the United States is more vulnerable than like, you know, random, you know, you know, and it's not clear to me, you get the same thing with AI, like it could just be like it just enables everybody to have bigger weapons, but the equilibrium is the same.

Well, I think that it actually has really huge impacts for borders and unfortunately I think China is well positioned here for a very specific reason, which is China, their, their justification for the great firewall is they've just justified it as digital borders. They say we can introdite physical packets, why can't we introdite digital packets? Right. And now with the whole Ukraine controlling drones in your territory thing, that becomes more than simply a metaphor, it's a real thing. It's like controlling cloud space.

Right.

And you know, if you can allow somebody to script drones or script humanoids in your jurisdiction, then they can blow things up, right? That's no longer a theoretical thing. That's, you know, and now the counter, counter argument is well maybe you just have them pre programmed autonomous so they don't even need an Internet connection and they can just do cameras or whatever. That's true. And in Ukraine, you know these things with the drones on cables, just crazy thing they're doing there, right? Like you know those like big unwindy cable things that you sometimes see on like ships, right. So they have these cables for these one way drones that are these ridiculously long, like multi kilometer long ethernet cables, like Cat 1 cables on a drone. So it can be Offline and it goes past signal jammers or something like that and then goes and blows up on its target and it goes forward and the cable gets tangled in the trees, they don't care because it's not going to go a reverse trip where it gets, has to yank the, you know, so when, so one way path, it doesn't matter. It's a one way drone, which is kind of crazy. So that's an argument that at least at short distance near the border, drones, you know, wouldn't be able to get, would be able to get anyone.

Right.

But that concept of digital borders becoming hard borders I think is going to become more of a thing where basically, you know, I, I, I actually had to give a Talk on this 12 years ago that you know, your immigration policy becomes your firewall because with telepresence you can move robots around and that's starting to become real. So that is something where, I think that has real implications for like the geography of a country. Because the alternative to that of having quote defensible borders is basically an encrypted state where you don't even know where it is on the face of the earth. And what I mean by that is can you make a map of Bitcoin? Not really. Right. It's something where it's so dispersed and you don't know every holder and they're moving around the world and there's no single map of every mine. And even if you got a map, you wouldn't know if it was complete or erroneous or out of date or something like that. You actually have in a sense security through obscurity. So you couldn't just go and blow all those things up versus something that's outlined on the map is recessile and vulnerable in a certain way.

Right.

So that's something I think about a lot in terms of like what do future borders look like. And you might have hard digital borders and China might preserve its territory, but those that can't enforce for rare reason, hard digital borders can't stop these kind of drone, you know, introgressions. Let me know your thoughts.

This is, I mean this, this, this assumes, this assumes drones are not autonomous.

Well, even at how, I mean they would still need to be given a control signal to do something. No, not on the third hand.

Can you really fully autonomous, you just be like, go blow up this building. Here's a picture. A fully autonomous zone would not need any packets.

That's true, that's true. And then also if you think about it, are you Going to block every, every single telephone call in every. I mean like it's really difficult. That needle in a haystack.

I have some super spooky stories of dig in China, man. I remember this is kind of a non secular, but I have to tell it was so spooky. I was in China probably 10 years now. I used to work for the government. I used to work for the intelligence community when I was a kid. Oh really? I was a kid. It was in 2001, it was my first job out of college. But anyway, so say 10 years later I was on a business trip in China and I called a friend of mine and I was telling about my day and the phone dropped. So I called my friend again and I was telling my friend about my day and the phone dropped again and I'm like, what's going on here? So what I was recounting is where I had been. So I called my friend one more time. I said 1, 2, 3, Tiananmen Square and the phone dropped. So, so I think there was a bug in whatever software they had and somehow I had been picked up. But I don't think it's too crazy to assume every conversation on every phone call can be monitored and has been for a while.

Well, so, so now, now they have something. This is another way where AI does change the balance of power in the following way. Right. Like in China, they, they had this saying which is the mountains are high and the emperor is far away. Right. And China always had a different conception of the balance of power between the government and the people than the west did.

Sure.

On the one hand, you know, and this is, this is a broad generalization over thousands of years history. But very broadly, like on the one hand, because the state had all the weapons and the army and so on so forth. Like, like that chair could morph into an agent of the emperor or a CCP guy today if, if, if the government so desired. Because there's no like limits truly, you know, in the sense of it can just do whatever it wants. Right. On the other hand, whatever the law is written down, the people just do what they want. Right. So they kind of, in a very pragmatic way say the limit is really the limit of what people can enforce. Right. And if the state has lots of power, then any written limit doesn't really matter. But the state can't find you since you're on the other side of the world. And then the written law doesn't matter either. Right. Which is a different conception than the progressive versus libertarian within the west, where they'll always quote law against each other back and forth. What is written is what is permitted or what have you. Right. But AI does change that balance because now the mountains are never high and the emperor is never full.

Long arm is incredibly. Because the long arm is infinite. Yeah, yeah.

They can synthesize. You know, there was something maybe, you know, this thing, Martin, I think it was called tia, Total Information Awareness in Iraq at a certain point where the idea was they had satellites covering a rock. And so every time some guy was put down, putting down like an IED or something like that, they would rewind the satellite to find who the guy was that did that and where he came from and then put a bomb through his window or what have you.

Right.

And in a sense it was like tracking someone for like their whole life, you know, because you were sewing together the trace of them through all of these cameras. Right. And that's totally possible for China to do now. And the difference is that AI makes it possible to. For a long time that data was ingested, but it couldn't really be parsed or queried because it was too difficult to look through five, you know, thousand hours of video on one person or whatever. That's increasingly becoming queryable. Right. And ingestible and summarizable in a way that it never was. And so I think that the real check on something like that is going to be going to have to be cryptography exit, you know, and so on and so forth, which is ultimately like get out of the jurisdiction, you know, have, have property that they cannot actually seize. Like you go back again to the limits of power and what have you. Right. Anyway, let me pause there, some thoughts on balance of power since you talked about that. Okay, last one, I think that there's going to be. There already is an anti AI backlash that's like the anti crypto backlash and will be part of the anti tech backlash. Because a lot of people are not using AI for what we're using it for. They're using it for like therapy or they're using it like, you know, for like as a companion or something like that, you know, and that's the top.

Of the pyramid of needs. It's kind of funny, like if you actually look, it's like self actualization, spirituality therapy. It's like finally computers are addressing kind of like further on up exactly the top of that.

That's right. And there's another aspect to it that I think hasn't gotten as much press, but that's interesting. To understand is that just like, you know, the tariffs are meant to kind of ward off Chinese competition. I'm not sure if they'll work. I, in fact, I'm skeptical. There's a similar, much less publicized thing that's happening at many media corporations where they're unionizing to try to ward off AI competition. Like they have union contracts that say editors, owners cannot use AI. Right. So making their organization very brittle where they think they own the market, but they're not allowing themselves to use AI. Right. And then eventually they're going to be beaten by AI enabled competitors that pull all of their followers and views and so on away from them because they're just more efficient. So I think that's going to result in an anti AI backlash. And I think that it's already kind of here where, you know, in some, on some like artist forums or whatever, they'll say, are you an AI supporter?

Have you heard that, you know, you know, that, you know, like AI artists spend as much time on building things as traditional artists. It's just a different toolset.

I know, but yes, yes, yes, that's true. But basically they feel that it's, it's kind of similar to the reaction by master craftsmen.

Yeah.

In the 1800s. Right. With when mass production started taking over what they were doing in the physical world. This is now happening digital world.

Right.

And so, and the other aspect of this is I don't think people have thought about the international aspect, where if you've got, let's say a lawyer who's making 200k a year in the US or a doctor in the west, and then you've got somebody from the Philippines or India or anywhere in the world who's making currently $2,000 a year. Maybe the converged wage with AI plus their IQ or what have you, AI plus human convergence is like 20k a year, which is a 10x for the person abroad, but a 1/10 for the person in the west, and it radically increases consumer surplus and so and so forth. But I do think that that's going to be something that's going to be, it's going to be a big deal in the years to come. And so we'll have to figure out how to mitigate that.

To your previous point, I just think this is so important. Like, I agree there's going to be a huge backlash and I think some of it's going to be rooted in like the experience of individual people. Like my job is shifting and that I am very sympathetic and I think we should Address. But I think there's something more pernicious going on, which is. And maybe this is my cynicism, but more and more, I kind of view politics as, like, you've got, you know, pretty sophisticated people and they have clientele classes, and then what, what they.

Yes, patron. Patron class.

What they say is, is basically what will move the clientele class the most. Like, that's what they just look for, sound bites that'll move the clientele class. And actually, you know, the, like, the, the, the patron are sophisticated people that can actually hold nuance in their head. They know they're complex topics. Right, Right.

But, but they dumb it down on purpose.

They dumb it down on purpose. Right. And like, what better talking point than AI? This goes back to the Promethean legend. I mean, we're terrified of technology. You can anthropomorphize it, you can talk about it as gods. I mean, it is the perfect tool to mobilize. And we're seeing this on both the right and the left. Right. So this is not in any way beholden to one party. So I think this is like the ultimate, you know, political, you know, tool for any purpose, and we're seeing it for that. And I think for that even more than crypto, by the way, I think, I think the EI strikes to the heart of people's insecurities more than crypto ever could. And so I think that this is the big battle. I think it's going to be bigger.

It's interesting. It's, it's all of the above. Right. Because AI is disrupting media. Crypto is taking power over money, robots are taking power over manufacturing, and drones are taking power of the military. So all of these. And by the way, there's a crypto angle to at least three of them because obviously a crypto angle of money. There's a crypto angle to AI in terms of constraints. There's a crypto angle to the drones because you're going to want the control plane for the drones to be on chain, since that's the part that can't get hacked. Whereas Pentagon Kitak. So I do think, yeah, there's, this is something where it's going after quite a few power centers.

300 years ago, if you said, well, listen, you know, we're going to do this new crypto thing, people would be like, but if you said, listen, we're going to create AI, these artificial intelligences that have unbound power, I think you're right. Really getting out of our human insecurity that we've seen in myth and legend for 3,000 years and probably longer.

That's true. That's true. I think. Well, we'll see what happens with, with, you know, currencies and so on. But I think you're right. It's both.

It's both for sure.

Perfect.

Love you, Martin.

Thank you so much.