#24 - Samo Burja - How to Prevent Civilizational Collapse

30:22In the 2010s, I well remember this feeling of excitement. We're about to figure out the human being, and who knows how our society will change, how marriage will change, property will change, how our ability to manage human violence will change, and all of this. But before we go too far afield from the initial question, I think that as soon as a society loses this ability to advance a living tradition of knowledge and is merely preserving just the pattern of behavior that produced the previous thing, it's like a timer. It's like your body has stopped making new skin cells, right? Right. Like I'm sure the only one you're just that you're just like, how long can you stave off entropy?

31:02Exactly. And that is a doomed battle that you can fight it for a very long time. Like maybe the aqueducts and say, I don't know, Constantinople in 300 or 400 or 600 AD, maybe those still work. But at some point, Constantinople is the only city that has over a million people in your rump of a Roman empire. Right. And Rome in the sixth century AD has 70 ,000 people and goats grazing in Colosseum. It doesn't matter if you're really dedicated to preserving the heritage of your city. In fact, late Roman emperors would sometimes put laws in place against people pillaging old abandoned buildings for building material.

31:39People did that anyway, of course. It's a losing battle. And metaphorically, what does it matter if we have something called a university, if a truly unusual thinker with a deep interest and knowledge cannot make a career doing unexplainable things for unexplainable reasons and occasionally showing up with a miracle. I feel like maybe in our quest to make knowledge legible too early, we sometimes destroy what makes it viable. We try to solve the succession problem through written tests, as if, for example, the many Chinese dynasties, like the Ming dynasty, had not attempted to do this. When Europeans encountered Confucian education, they were first immensely impressed because it seemed very close to their ideals.

32:30If you read 18th century Europeans, they find immensely inspiring the Confucian educational system. People like Voltaire hold it up as an example of a rational civilization, a wealthy civilization. In the 19th century, they complain about how it's all rote memorization. I think, honestly, they're both right. In the 18th century, that system still works very well. By the 19th century, that same system, as you run its internal logic and as everyone follows the incentive gradient, people come to memorize the leaked test results. And studying for the test at first is a great boon because at first you're sort of strip mining all of society for talent.

33:11And the people who mastered the test from any layer of society, they're very good at mastering anything, really. They're just very smart people. But once that process has been running for a few generations, I actually think it's inevitable that the imperial exam or the civil service exam, or let's say the entry interview, the acceptance interview for a Harvard position, becomes a farce, becomes something where maybe, I don't know, your entire high school experience is choreographed to look good at that one interview at an Ivy League university. I think we're already there. Yeah, but we're most certainly there.

33:47We are certainly there. And I think we were not there 100 years ago. And we were not there 50 years ago. Okay, so what are some other signs that you've noticed that suggest that we are starting to mirror some of these problems? Well, I think that it is very telling that we have found a general disappointment in the rate of advance of most technologies outside of, let's say, anything related to the information revolution, anything related to the AI revolution. So this is the classic sort of Peter Thiel observation about bits versus atoms. Currently, the promise is that in the very near future, AI will revolutionize atoms.

34:30For now, I'm only seeing it make words. It makes words very well, makes them definitely at the level of a college undergrad. But perhaps that's an indictment of the college undergrad. As much as it is a triumph of open AI and anthropic engineers, I think that the fact that we had millions of people basically wrote essays, essays that can be statistically predicted, just shows us how deeply we exhausted what was once a deeply intellectual exercise. The essay comes from the term to attempt something. The idea behind an attempt is that you put into words and put thoughts on paper that you've never explicitly examined before.

35:18Who can say that they truly wrote an essay of that type in high school or even college? I can point to my, in all my writing, I think I have something like six articles that truly achieve that level. And for many of the other articles, I'm just explaining what I learned from other sources, or what seems to be the ultimate evaluation. And I think the fact that we were a writing-based civilization, and now we're sort of almost LARPing being a writing-based civilization suggests to me we have lost huge swaths of tacit knowledge. I would even argue that the stylistic differences between scientific papers in 1920 versus 2020 suggest that we have already lost much of what was once the practice of science.

36:04Well, read a scientific paper from 1910, 1920, 1930. Pick a topic of arbitrary complexity. It can be a quantum mechanics paper even. the style is going to be much less dry than any modern paper even any modern paper it's going to be less formulaic it's going to be much less formulaic it's going to be very economical with jargon it's going to be very straightforward very precise but also beautiful why is it beautiful it's beautiful because it's meant to be read by human beings not committees it's beautiful because it is a remnant of the old peer review which we have nearly forgotten the old peer review being, hey, I trust this person's intellectual taste, I will submit to their journal, and they will decide if they publish me in their scientific journal.

36:53If my work is bad, I'll be embarrassed in front of our scientist friends. That's old peer review. New peer review in the 20th century is the nominally meritocratic and rational committee. The problem with the committee is you must agree with every member of a committee, even if they disagree with each other. So over time, this has produced a very stilted form of writing, and it is a writing that is dreadful to read, right? Even if you're super interested in the subject matter, which means most academic papers are not read. So who are they produced for? Why are we producing so many papers? Well, it's hard to give the government a KPI for the scientific research it's funding.

37:33An easy KPI is, oh, we have 50 researchers working on this. No, we have 500 scientists working on this. No, we have 5 ,000 PhDs in chemistry working on this problem. So the inflation of headcount. And you can also say our department has published so many papers, so many scientific papers. At one point, the scientific paper was obviously a good advancement. Now, let's note in social science, for example, we had the replication crisis go through and basically torch entire fields of knowledge. The publication of a paper is not news anymore. it's like not even news for people in the field right but it is a requirement to publish papers to achieve funding and to achieve accreditation so that's this hyper production of low quality writing where the quality of the writing itself no longer matters betrays that we are no longer a writing based civilization we just believe it's super legitimate to base authority in writing So it sounds like what you're getting at here is it's a kind of, it's like a dark amalgam of excessive bureaucracy and like tyranny of metrics in terms of optimizing for those.

38:49Goddard's law in everything. Exactly. Goddard's law. I mean, once an excellent measure becomes a target, right, it ceases to be a good measure. Exactly. And so everything has become, you know, before your goal was writing good papers to do good science. and then it's like okay but now the goal is to just write papers and and so you're just your objective function has like gotten derailed um and you are and and yeah so everyone is now sort of trapped in these i call them moloch traps but these these kind of social dilemmas where even if you don't want to be churning out just paper after paper to get funding if you don't do it everyone else is going to so you have to play that game as well you are stuck in that equilibrium no matter what you want to do as an individual, unless, for example, if one were independently wealthy and very socially resilient, then one could fund any research they wanted or sustain themselves.

39:41But very few people do this, right? It's one of the reasons I say admire like, you know, the Stephen Wolframs of the world, whatever you think of the scientific merit of some of his theories or the self-presentation. I actually think he deserves praise for trying to actually be a scientist in the way a 19th century scientist would have tried to be a scientist, right? You know, we forget the idea of the sort of gentleman scholar free from worldly affairs just trying to figure out the world, right? And sometimes having a big ego. Isaac Newton had a big ego. That didn't stop anyone from taking Isaac Newton seriously.

40:17would Isaac Newton have done well in any sort of HR department interview? Horrible. He would have done horribly, right? He could be positively nasty. It wasn't just like Asperger's, right? He could be viciously competitive. And I think we have come to deeply disdain the kind of personality that can actually drive new ideas. I think we have put a high premium on being very well socialized, by which we mean very conflict averse on the surface, but perhaps very conflict oriented below the surface. In our society - Right, that energy's got to go somewhere. It has to go somewhere. In our society, any sort of physical violence, of course, is out, which sure, okay, that seems reasonable.

41:11But even overly harsh scientific criticism is met with suspicion. Why are you upsetting the apple cart? Shouldn't we all be working together as a field? Blah, blah, blah, blah. What are we doing? Are we collectively preserving class privilege, the class privilege of being an academic, or are we trying to discover truth? Or economic competition, trying to do economic competition for real today will get you denounced as a ruthless businessman. When it's like, really, well, all I'm trying to do is make a good product, and we are going to try to drive our competitors out of business. If someone says that in modern corporate culture, they're like, what, top four percentile, top two percentile disagreeable among CEOs, right?

41:53Maybe top 1 % even. These are all of these ways in which, of course, the fact that all of these things are punished, right? Physical violence, intellectual competition, economic competition, that of course suggests there's a lot of hidden physical violence, intellectual, you know, intellectual sabotage and economic sabotage. because otherwise who would be punishing them, right? So there's a broader theme, I think, of we have focused very much on eliminating outliers and we're experiencing kind of evaporative cooling in society, right? Where over time we always lop off the right and left side of the distribution or maybe just the right side of like a Gaussian distribution.

42:39And, you know, not in a genetic or IQ sense, just on any sort of trait that we allow people of various positions to have. And over time, our sense of what the proper social role is gets narrower and narrower and narrower. And we have, in the process of narrowing it, cut out huge chunks of what once made it functional to be, I don't know, the member of a corporation, right? A functioning board today is exotic, almost dangerous, almost unprecedented. like a board that practices any oversight over a company that's that's weird right like how everyone had a meltdown over that when the open ai board i mean its job is to keep an eye on the ceo and while none of us really know what actually went down like they clearly felt strongly enough that the ceo had done something wrong and they took an action and then the whole world had a meltdown about it and it's like that's what a board is meant to do so the fact that that was so shocking suggest that like something's gone awry there we expect a certain level of um well what do you think like it was that was that a further example of this or was that actually something against the trend you know i think all indicators suggest that sam altman is an excellent uh ceo according to most metrics used in silicon valley however it's undoubtedly true that that is exactly within the authority of a board right a board might be wrong but the fact that it's become so rare to have any sort of board oversight in a modern corporate body in Silicon Valley, even an unusual one like OpenAI.

44:13I think that's concerning. So even if the board was incorrect in that case, right, it still suggests we've lost the social technology of a joint stock company. We've lost the social technology of a well-governed nonprofit. It means we've lost that feedback mechanism, basically. I mean, it legally still exists. All of these things exist on paper. As I'm sure you know, the Roman Senate existed on paper long after it ceased to be relevant, right? In the 7th century AD, in Constantinople, you still have people going around calling themselves senators, right? Would they meet? No. Would they pass laws?

44:47No. On what stage of this change are we, say, in our political institutions? I would argue that the majority of law in the modern United States is made purely administratively. What is not made administratively is made through judicial review, and only a small amount of law is made by Congress itself, right? So when people are saying Congress doesn't do anything, yet then turn around and speak about our democracy, that's very strange. I think this will perplex future historians for much the same reason we are a bit perplexed that Augustus is calling himself first citizen rather than just he's obviously the emperor, okay?

45:25The Senate obviously doesn't matter and everyone knew the Senate doesn't matter, but they drew legitimacy from long dead institutions, right? This is also another thing that happens. Over time, the maintenance of the appearance of an institution can become very vital for the political balance of a society. So you have zombie institutions going around where we need someone to fulfill a social role, and the role of a long broken institution is to stay there and pretend to be in that social role. And, you know, I could say quip that maybe the modern United Kingdom is a bureaucracy pretending to be a democracy, pretending to be a monarchy, right?

46:07Where, you know, any sort of democratic element, when it does have input, it's very unconsidered, very chaotic, maybe like Brexit. And, you know, whenever any monarchical element enters, it's this very strange, untouchable thing that Haber is still there and will probably still be there 200 years from now. I would say that I don't think the UK is pretending to be a monarchy anymore. Like no one in, no Brit thinks that the monarch actually has any meaningful power or very few Brits do at least. But I think they do, I do agree with you that it's at least they still hope and believe that democracy is legitimate and playing out.

46:49So what, when you say it's a bureaucracy masquerading as a democracy, what are the signs that it is, what makes you think that it is no longer a functioning democracy? And was there a particular tipping point where you were like, okay, this is where it switched over? When was it last, basically? Well, I would say that a very important point has to be that we all basically explain away the failures of elective representatives to institute any sort of change in purely personal terms. We always talk about, oh, these politicians are corrupt, bad, not serious, co-opted by the system and so on. Yet, obviously, anyone who makes it to those positions always has those features.

47:39The system and the whole set of incentives actively select for a certain kind of individual to go there. But the pretense we all have is that if the right individual got in, they could still do things. Why do we believe that? Why do we think that if suddenly a different person came into parliament or Congress, that they would be able to organize with other fellow congressmen or members of parliament and say, I don't know, reform the intelligence community or reform the Pentagon or even reform either way the NHS, right, in Britain, I think it's very dubious that they could do that, especially in the very short amount of time they have available, especially with very limited resources in terms of even research capacity.

48:31Like how much research capacity as to all of these difficult problems that an elective representative is supposed to have a crisp opinion on, how much capacity they have. Who is doing that intellectual labor for them? Well, might be lobbyists, might be their political party, if they're in a particularly functional party, but they're certainly not doing it as individuals. So I'd put the tipping point around the last time you saw a representative, deliberative legislative body where a speech was given to other members of the elected body with the primary intent to change the opinion of fellow legislators rather than the general public.

49:13Now, here's an interesting one. We're used to debates being a circus. They are there to help you be elected. It is exotic and strange for us to imagine a body where the elected officials meet, and they try to persuade each other of points. Right, and actually change people's minds. Exactly. Truly. Yeah, they're trying to change each other's minds to converge on a common truth. Conversion on common truth with each of them being a representative of the interest of their constituents. What I said sounds like civic 101 because it is. And because it's crazy to point out that's not how it works right now.

49:48But the crazier thing is it kind of did work that way in 1900, 1920, I would say as late as 1960. I think in 1960, if you had a persuasive, well-argued speech in British Parliament or in Congress, people who at that point were 70 or 60 years old, who were not in touch with the incentives of their era, they would probably behave as a real deliberative body, right? And I think in the United States, an interesting point is the 1970s, which is when Congress, to some extent, attempts to both rein in the president, but also rein in the intelligence community. There's commission hearings that declassify a bunch of stuff in the United States after the Vietnam War.

50:31Of course, the impeachment also of Richard Nixon. But that's sort of the last spasm. So if we were looking at US history, I think, you know, the 1970s are a good mark. And for the United Kingdom, maybe 60s, maybe 70s. What role do you think the sort of sensemaking crisis that we are increasingly getting into could be playing as well? Because again, you've talked about how succession of knowledge, the continuation of tacit knowledge is so important. It's arguably harder to actually get a clear picture of the world now. I mean, in some ways, yes, we've got more information available than ever before.

51:15The world is more connected. But at the same time, we've got generative AI that's going to be very quickly making very convincing politically sensitive videos if people want to, you know, troll farms, etc. And then coupled with the fact that there's all these different sort of incentives at play to encourage newspapers to skew their headlines in favor of their existing readership and so on. So it feels like it's getting increasingly hard to get an accurate picture of the world. that seems to only go in one direction, right? I think that absent certain important countertrends, we are seeing over time the collapse of our society's capacity for collaborative, for basically collaborative sensemaking in the sense that you describe, right?

52:14We are seeing over time a breakdown of our ability to process information collectively. So we used to greatly rely on a relatively healthy epistemic commons. We used to rely on many incentives that kept, let's say, the quality of the median book above a certain level, or the quality of a certain broadcast above a certain level. And over time, we have fewer and fewer incentives to make good faith contributions into the epistemic commons rather than try to, in a variety of ways, capitalize on the epistemic commons. exactly like we we might want to increase our ad revenue we might want to um you know pursue our career and the development of our career even if it means a certain amount of intellectual dishonesty in an academic context uh we might make decisions with regard to algorithmic feeds where the information presented to the user is not the information that is most useful to them or best for them, but the information that keeps them checking their app, right?

53:28So these many small decisions together produce an erosion of our common body of knowledge. I think one of the beautiful things about generative AI so far has been, I almost view it as this very beautiful last monument of a great civilization, where the global body of text, once you run through it, Once you run the sum total of all the text that's been produced so far, you get a pretty competent undergrad writer. That's how I would describe like a chat GPT or a Claude. They're actually like balanced relatively. They sometimes make stuff up, but on that they write not too badly and they're getting better and better being accurate.

54:14But in a way, they're giving us the gold standard best answer from this like corpus of human writing, right? I mean, you could argue that it's the models. There's still scope for the models to improve. So yes, the corpus of text is fixed, but... The models might improve. The models might improve and get better at inferring what you really mean, etc., or drawing threads of... My claim would be the quality of the corpus of text is going to decline. Right. Well, especially as I'm hearing that because now all the data that is available... I mean, probably not all books, let's be real here. not all written books are now digitally have been digitized yet so there's actually a large repository of very high quality text there that arguably should be fed in um and it hasn't yet however that's a one-time game that is a one-time game yes but i mean it should be a quite a big step change but what i've been hearing is that companies are now looking into synthetic data and that sets off a lot of alarm bells because basically that okay we've run out of internet to scrape let's come up with more synthetic text in which to train our models further but it's like isn't that like garbage in garbage out in a like horrifying feedback loop i mean again i'm sure they're thinking about this but it it seems yeah if there's like a ск.se we need better ways of sourcing meaningful information and knowledge as if we're going to be feeding into these AIs that are going to be useful for us.

55:41But yeah, that's an interesting way of putting it as a like last monument. I mean, it's a summary, right? We have written the Sparknotes on our civilization, and we can query Chachapi to get the Sparknotes version of our civilization. TLDR did not read, right? The TLDR of our civilization is in every answer to every prompt you might ask of a modern generative AI. And, you know, the question is maybe the TLDR is good enough for us to get by, but to contribute to that corpus, the continue either the scientific or the philosophical written tradition that doesn't exist. And because of that, I mean, possibly it's truly obsolete, possibly video, possibly, you know, generated AI processing can substitute all of those knowledge transfers that used to be oriented around a practice that had written output.

56:35But I doubt it. So for that reason, I expect the breakdown of law, right? Law is perhaps one of the last remnants where we are a writing based civilization. I talked about the breakdown of science. We could talk about the breakdown of cultural instruction through writing. Arguably, this is one funny way in which we are less Protestant and more Catholic. So in the 19th century, if you had emotional problems, you would try to read an insightful novel about the human condition. Or you would try to read material to help you regulate your feelings. So like Dostoevsky? Exactly. Something like that. The novel was super popular because it was a psychological aid.

57:16We talk about entertainment, but really what we should be talking about is self-medication with information. So that's why it was super popular, why novels could be even addicting. People talked about addiction to novels in the 19th century. And in the 20th century, we discovered that confession works wonderfully. So we secularized it and we called it therapy. So you have a high status person and they talk to you about your problems and you give them some tribute and this relationship is sustained. And so that's my joke, right? We went from a book-based civilization like Protestant to priest-based civilization, Catholic confession, secularize this therapy.

57:57But anyway, let's say that our emotional self-regulation has somewhat broken, whether it's because of new developments or because old technologies just don't work anymore, old social technologies. That's an open question. And then I talked about how economic competition is considered almost unfair in a deep way, right? Even, for example, you know, and there are problems with doing business with China, but I know to myself that TikTok definitely disproved the narrative that Chinese companies could not make compelling apps for users or compelling software for users. Now, maybe we still ban it here in the United States, but I think the fact that TikTok was successful really spooked many people in Silicon Valley.

58:44And the first impulse wasn't, we're going to build an even better app to displace TikTok. The first impulse was, well, this comes from a communist country. Surely their government has its fingers all over it. Let's ban it. Like that's, I saw like, I had hundreds of conversations where this was sort of the first impulse. But I mean, there was evidence to be fair that the app has all these like backdoors. Sure. Do American apps have all of these backdoors? Supposedly fewer than TikTok. They were finding that it was scraping your clipboard of your phone, for example, copy paste and so on. Why is the solution to that not a better American TikTok?

59:23Sure. I mean, I would argue that there's a reciprocity problem with China. China's banned all the American apps from its app store. Why are we allowing a Chinese app? Well, not all of them, right? Like Amazon could operate in China. Sure. But the point is, it's banned all of the social media ones, for example. So I think that unto itself, like you do need to have some degree of reciprocity. I'm not even arguing that TikTok should not be banned. I'm just saying that there are signs that we are uncomfortable with economic competition. And there are, politically speaking, I think sort of law currently remains this thing where it's become ever more load-bearing, right?

1:00:06The US is an ever more lawyer-run society. We could joke that America is a word-sell society. We use that kind of dumb meme. And in some ways, it's like by relying more heavily on law, it's kind of the equivalent of running to teacher. It is running to teacher, but it also means all the other systems have broken down, right? Because it's the last remaining thing you can go to. It's the last remaining thing. And then if ever we have generative AI, excel at law, we will have a wonderful period of cheap and high-quality representation for everyone, followed one to two generations later by a much reduced human legal tradition.

1:00:49So in a way, we will fossilize it. Well, because that service will displace the economic niche of a lawyer, and eventually the socioeconomic niche of a Supreme Court justice. Can you imagine Supreme Court justice that has not had 60 years of reasoning in their own head about legal arguments, but in fact has 60 years of clicking yes on a suggested text. Horrifying. Again, lost task knowledge. Exactly. And even if it's not the Supreme Court justice, if it's their clerk, that's the same thing, same thing, really, even if it's their junior research staff. And again, the result is sort of like perhaps eating the epistemic seed corn.

1:01:33Now, possibly again, eventually the AI is just good enough to be some sort of virtual person in our world. Maybe we're comfortable outsourcing law to artificial intelligence. But at that point, it's really not our civilization anymore. At that point, in a way, the enlightenment has ended, right? The project of the enlightenment was supposed to be human reason, exploring society and re-engineering the systems and incentives for this kind of the benefit of all, right? That was at least the pretense of that project. And all of these things like representative government, the scientific method, even market competition, and at other times enlightened central planning, these were all ideas coming from that philosophy.

1:02:24And we return now to the haunted world the world where the machine says no, where the machine says yes, and we don't really know who builds the machines. We don't really know whether they work or not. I'm not even horrified about the potential of an AI takeoff or AI actually being in charge. I'm horrified of us becoming so superstitious about our machines that we develop a priesthood of the AI long before the AI is actually capable of running our society. Treating it as an oracle when it doesn't deserve to be. Our minds might use good enough year after year, even if good enough gets worse and worse, right?

1:03:07Like you can have a very cheap, low quality good can often displace, you know, a higher quality, more expensive good. That's okay. But when it comes to the epistemic commons, there's like fewer and fewer reasons to do a good job. And it's not something you can legislate, not really. I feel like legislation at most is like putting band-aids on this wound where we're bleeding out a well-maintained commons that once was just honestly maintained by people who inherently cared about intellectual integrity. It's hard to remember, but we used to care about intellectual integrity in this almost religious way.

1:03:46The important thing as an individual in an enlightenment-style society was that you cultivated these virtues because this is the means through which the world gets better. So you would do it even when it wasn't economically the best thing to do, when it wasn't the socially best thing to do. You did it because you basically believed that this is how we, humanity, make the world better. and the step of we will remove this and put this responsibility on the machine is a very, in a way, reasonable one. But as soon as that's done, all of these reasons to maintain your ability to process information of the world, they go away and the whole house of cards collapses.

1:04:24And then you have far fewer people who have this built up machinery, this incentive to do this. And I claim that will feed right back into diminished manufacturing capacity and actually slower progress on fundamental algorithmic breakthroughs. Like every time I talk to someone who makes on me a strong impression that they're going to drive AI further, it's never the sleek marketing guy. It's always some Israeli or ex-Soviet mathematician weirdo that has strange theories of metaphysics, strange theories of math, and strange bets on AI. When I meet someone like that, I know that I met someone that might bring AI forward.

1:05:05When I don't think I've met someone that will bring AI forward is when they talk about it like it's a plane delivering cargo. Right? The literal cargo cult. We are very close to saying that we don't need to know how to build a much better artificial intelligence. The next model will deliver the cargo. It's going to bring us nice things, and we're just going to enjoy them. No, no, I'm talking here about the cargo cults where, you know, the assumption is that material goods come from, you know, planes filled with U.S. soldiers. Right. So it sounds like what you're getting at is there's this kind of ossification of society.

1:05:48The simulacra of reason is beating out reason at all levels, right? At the most elite levels and most popular levels. So is this just an inevitable process? Is it like part of the cycle of every civilization? Have there been any that have broken the way out of this cycle or at least gotten close to it? I think that a certain loss of particular traditions of knowledge is probably deeply behind almost every civilization that has ever collapsed. And the reasons is it is always outcompeted in a certain sociopolitical game, right? So, for example, having an excellent understanding of the Roman legal code at one point was an excellent way to achieve power in the Roman Republic.

1:06:35That becomes obsolete when the correct way to achieve power is command of a Roman legion, say, crossing the Rubicon. right? That becomes obsoleted when the correct way to win power is no longer not even being a good military commander, but it is winning court entry games and assassinating any general that happens to be too successful. We could name Belisarius in the empire of Justinian, the Eastern Roman Empire, or commanders in the 4th century AD under Emperor Honorius, right, who would just kill successful generals. Because why? A successful general can be a threat to the supreme power. So we have a certain time in almost every civilization's history where a particular skill or a particular tradition of knowledge, be it military, economic, scientific, spiritual, you know, social, right, urban, where developing it is the best way to achieve success and recognition of society.

1:07:40And the tipping point of diminishing returns comes sooner or later, right? When the S-curve is exhausted. And at that point, you go from positive sum games to negative sum games. And as soon as your civilization starts playing negative sum games, that is infectious and spreads rapidly through an institution. If the best way to succeed in a major government bureaucracy today is to treat it in bad faith, if the best way to succeed in the media space is to treat it in bad faith. If the best way to succeed in academia is to treat it in bad faith, wow, what's left? And I'm not even done. I could go further.

1:08:19If the best way to run a startup is to treat it in bad faith and the best way to do it is to hype it up at all times, even if you know your technical fundamentals are sketchy, then maybe we'll all become, in Silicon Valley, we will all become Elizabeth Holmes. Maybe we've already become Elizabeth with homes, right? And we don't even know it because who would point it out, right? It's just so default. It is very default. And that's why I think that how to break out of that as a civilization? Well, sometimes you can take very extreme measures, which I think can prolong the existence of a society, right?

1:08:52The creation of a completely new city that comes to be a new elite center is surprisingly effective. Why? Well, human beings are biased against moving. Most Roman noble families did not want to follow Constantine and his weird bunch of religious freaks out into a random Greek city, formerly Byzantium, then Constantinople. Yet even 50 or 100 years ago, they found themselves radically disempowered compared to the new senate that was set up. There was a separate senate set up in Constantinople and a new military apparatus and a new taxation apparatus focused on the eastern provinces. And of course, the pagan state-backed religion gave way to the Christian state-backed religion.

1:09:38So that was a transfer of power there. There wasn't a direct fight. There was just this sudden change of where coordination happens. I think cities need to be understood as physical embodiments, instantiations, incarnations of coordination. Which street you live on, shockingly important. Which offices are close to each other, shockingly important. When the U.S. was like structured a bit more honestly, there was a, you know, there was a department of war, rather a department of defense, and it was in the same building as the State Department. I guarantee you, if your generals and diplomats run into each other in the hallway, you will be in fewer stupid wars.

1:10:18As soon as you physically separate them to opposite sides of a river or opposite cities, the left hand doesn't know what the right hand is doing. And actually, you might have basically exiled someone to a place of irrelevance. It's not just Roman history that has this phenomena. Russian history has this phenomena. When the Tsars wanted to be absolutist in Russia and have great power over the high nobility, they moved out of Moscow to St. Petersburg, leaving the Moscow nobility behind and irrelevant. So I think I would be very optimistic if any major Western country today created a new capital with a new university with a new special economic zone.

1:11:04I think I would be extremely optimistic because that would mean that the old negative-sum institutions are left to play their negative-sum game, which is self-marginalizing, and you immediately are starting something that at least locally in that city can be a positive sum game. A new flower blooming. Yeah, I mean, this maps very closely onto what Maghut Wade, who was a recent guest, was talking about. She is trying to get these, she calls them charter cities. And presumably this is, are you referring to literally like building a new city from scratch or is it like repurposing a new city and saying, look, we're going to try out our new set of laws?

1:11:35Honestly, it's good to have a new city from almost from scratch. Though it can be an existing city if a completely new industry is instantiated there. Any potential ones in the US do you think that could fit that bill? I think the United States is... Too calcified? No, no, no. I think it's actually fairly lucky as far as Western countries go. I think the demand for urbanization here outstrips supply. There's still like 50 to 100 million Americans that if they could move to a big city and have a high paying job would happily do so. I claim this is no longer true of say Italy or Spain, right? The countryside's already emptied out.

1:12:12The second tier cities are emptied out. Slowly, even the first tier cities will start to empty out, say like Madrid or in Catalonia or Milan or whatever. Empty out to where though? Just population decline and also to even higher tier cities, right? Why live in Madrid if you can live in Brussels? If you can live in Brussels, why live in Brussels if you can live in London? And why live in London if you can live in San Francisco or New York or Texas?

1:12:44right so the the other method short of creating new cities which i think is a very it's it very much gives you like a sort of a societal reset uh the other the other possibility is uh pursuing a new field of development so if your society had basically been occupying its elites through like a religious system, for example, opening up a commercial way for elites to pursue games, right? So people who are very ambitious, who are very skilled, if they come into an environment that is a positive sum, like a learning environment even for them, if most of someone's career is positive sum, well, they will develop positive sum skills.

1:13:25If most of someone's career is negative sum, they're going to develop negative sum skills. Who do you wish to be surrounded by? So if maybe, say, an extracted, somewhat decadent, overdeveloped religion, such as some people would say the Catholic Church was in the late medieval period, encounters a vibrant, meritocratic, bureaucratic sort of state where it's a young bureaucracy, not yet dysfunctional, like, say, maybe the French state was or the British one, power very naturally shifts to the other one, and civilization as a whole changes, but it becomes positive some game again. And then maybe later on, say, religion as it persists in the 18th or 19th century becomes positive sum again, because they are no longer the predominant trendsetters of society, and they have been sort of emptied over time and starved.

1:14:14The negative sum game is almost cauterized into this field that is ever more marginalized. And at one point, it's flimsy, and the negative sum sort of games are easily overturned by new positive sum games. My hope had been for almost the last 10 years that tech would prove to be the positive sum game of our civilization, which forged an elite trained on positive sum games that would then go and, I don't know, revitalize legacy industry, you know, the manufacture of atoms, or maybe revitalize government, or maybe revitalize culture and media. And while we see a few excellent examples of this, we're not seeing anywhere near enough to overcome the extremely negative trends in any of the branches of government, in any of the major professions and industries outside of tech.

1:15:07Do you think that is because the game that is developing the tech itself is too already malachian and negative sum? I think it was very positive sum in a lot of ways. But again, it's still beholden to a lot of these misaligned incentives, right? They're still optimizing for, well, we've got to be profitable, we've got to raise our next funding round, who's raised the biggest funding round as a lot of sort of dick swinging and so on. Is that the reason why tech is insufficient to write this? Or is it just because it's like slapping a bandaid on a rotten, like a festering wound that is much deeper, like at a systemic level?

1:15:44I think that to a great extent, tech is very inward looking. It does not look at the rest of society very much. I think if it looked at the rest of society, they would just choose to do many things very differently, many things that are easily preventable, right, easily can be improved upon. well let's just go for for a little bit and acknowledge that back in 2009 or 2010 the search engine was a wonderful innovation that lets you think about the world in a completely new way and honestly if i could choose today between having a 2010 era search engine or a 2020s era generative text model, I would choose the search engine.

1:16:32Really? Yeah. I think currently search engines are terrible, but they've become much, much worse than they were 10 years ago. The median quality of the information and source I could find in 2010 was very high. If I wanted to find a high-grade crackpot autist who was completely correct about some small niche thing in 2010, that was easy. I believe those still exist, but if I try to find one today online, I don't find them. And how much prompting would I have to give to a model to have it say, I don't know, replace Gwern, right? Gwern, Gwern.net. It's impossible. I cannot prompt the LLM to give me Gwern.

1:17:13I could find Gwern in 2010 Google. Right. If you're a new person, you're not finding them now. You're not finding Gwern in 2024 with your search engine. So, okay, there was a massive positive information revolution in the 90s and 2000s that actually counteracted and corrected for somewhat the decline of academia. Without the internet, it's not clear to me at all that the replication crisis would have even happened. I'm not even sure we would have figured that out. It wouldn't have been a crisis in the first place or we would not have? We would not have diagnosed it. We wouldn't have diagnosed it.

1:17:46We would just believe the bullshit studies. And we have believed it for decades to come. which is a thing, nature allows that to happen. There's nothing just universe that prevents us from being deluded for a century or more.

1:18:03But you challenged me to say what could tech do better. So I'm going to give as concrete an answer as I can what tech would do better today after having acknowledged immense positive contributions. Like Wikipedia again, it's not only a TLDR of everything our society has achieved. It was actually very productive in the Enlightenment sense, because people would use Wikipedia, reason about it, and the community of Wikipedia editors, which, of course, they had their own political biases, but they maintained rigor and excellence for basically social incentive reasons. So in a way, Wikipedia was closer to old peer review than bureaucratic new peer review is, right?

1:18:41So I think Wikipedia is a great example of technology entering our society and picking up a function that, you know, 200 years ago would have literally been down to the editors of encyclopedias who at the time thought that they were transforming society by giving everyone access to knowledge. And in a way, probably they were. still acknowledging all these we could talk about 10 other positive contributions but acknowledging those what can they do today you know I think the best things for tech to do unfortunately are not the most profitable I think no matter what you do the margins for software are always better than hardware we have saturated much of what can be done with software we should be pursuing many more hardware improvements.

1:19:31At the very least, I would wish to see top tech talent allocated with significant funding to optimizing the factory floor the way we have tried to optimize the news feed. Can you imagine a big data approach to manufacturing? That's real automation. That would be remarkable, right? I believe further that a lot of tech wealth could be deployed in a very classic way that is not understood very well, which is socially responsible philanthropy. I think we live in an era where there is no for-profit reason to give accurate information. I think what should be done, and ideally, the billionaire should be diverse in their thinking.

1:20:21I think if every billionaire basically funded a private blog with their favorite thinkers, and they took pride that these are different and distinct from what all of their friends are funding. I think those could be very positive nonprofit organizations. Keep them small, keep them diverse, have them be these sort of like, you know, focused on excellence, have them critique each other basically for social status points because the monetization just seems too tricky. I think that could work extremely well, especially if one cauterized the sort of like new york media scene you know by hiring former and current new york times journalists that's how you go from being early wired to being late wired you mean the magazine the magazine yeah yeah per se i don't know mike solana's pirate wires is as soon as they start hiring professional journalists it's over knowing mike i don't think they're going to do that he's he's got i mean but at the same time the the incentives are there right you want you want the big names so yeah and we could we could go further and i think that i'm glad that say gary tan is undertaking some amount of uh let's call it uh you know yimby activism in the bay area probably you know if you're uh leading yc you correctly identify that Bay Area rents and lack of construction are actually eating up a lot of the money one invests into software companies.

1:21:51But I really do think that a new look at urban politics in the United States would do much to fix the problems of society. And it's somewhere where I think technocratic solutions just would work. So again, the talent, the IQ of the industry is there and tools could be leveraged, right? Information tools. Like what though? what would be the technocratic solution to these policy bottlenecks? Literally spend any effort at all to try to change the policies. So in other words, I'm saying, I wish tech elites were as politically active in our society as journalists would like to pretend that they are.

1:22:30I see. I see. Given they're getting called that anyway, they might as well do it. They might as well do it. There's no negative to the reputation. In fact, maybe they are actively punished because they're not doing it. Who do you critique? You critique the person where there's no consequence to critiquing them at all, right? And with regard to, you know, with regard to the, you know, leveraging of the unique expertise, right? I think the space of organizing voters, even the voters themselves being organized, right? So not even top-down organization, and just giving voters tools for them to organize themselves.

1:23:07Massive disruption opportunities. People have totally stopped talking about digital democracy and stuff like that that was very popular in the early 2000s. It's gone. Mostly because you can't really easily make money off of it. But the positive returns to fixing a country's politics would be immense, even eventually financially. If you live in a well-functioning country, so many things are much easier business-wise. You wrote this book on great founder theory, which I'll try and paraphrase is essentially that some of the greatest civilizations or institutions in history were off the back of either a singular or a small number of incredibly talented, forward-thinking individuals.

1:23:54And that was an integral part of sort of our march, humanity's march of progress. what role do you think that that could play in trying to right the ship of current western civilization well um you know to uh to restate great founder theory as i formulated it it is uh you know a variant on the so-called great man approach that historians in the 20th century critiqued in favor of trends forces and so on and uh the key the crux of it is that i believe innovations in social technology, that is human coordination, and that is human organization, are usually the work of an individual or very small groups when they're first prototyped and introduced to society, but once introduced to society, have a footprint that can last thousands of years.

1:24:45And very simple examples would be, even we talked about Confucian civilization, Christianity, all of these things. Well, organized religions usually have a founder, right? Sometimes it's not the founder it's credited to. We could debate about the true origin and development of, say, something like the Catholic Church, or something like the sort of Confucian strategy of, we're going to study the classics, and then we're going to travel around the warring states, period, China, trying to convince the princes of these states that they need very high-skilled civil administrators. And then we're going to run that little statelet well, and eventually that statelet is going to be all of China.

1:25:23So, and by the way, that is literally sort of Confucius's life and the life of his students. They were just these wandering scholars trying to explain why they would do a good job of running things. And shockingly, it worked, right? You wouldn't think that you can have, you know, sword-wielding rulers be talked out of running things. But it's only when you remember how tedious it is actually to run things where you realize how easily that can work. Would the founding fathers of the US be an example of great founders? Yes, very much so. I think that it is hard to deny that they're definitely one of the, you know, peaks of sort of Enlightenment era civilization.

1:26:00Their aspirations were not limited, right? If you see, you know, Jefferson, Thomas Jefferson, like editing the Bible to be more rational, or Benjamin Franklin being very interested in, you know, science, natural science, That was not unique. That was actually common in their milieu. And they definitely debated for real and wrote pamphlets not just to persuade the public, but to persuade each other. And they had vigorous disagreements about - They actually wrote essays. They actually wrote essays. They were discovering. and they gather together. And if I told you that we're going to gather 300 of the smartest and most powerful Americans today, and they're going to decide what the next government will look like.

1:26:41That sounds a little bit scary, a little bit exciting, a little bit impossible, but that is literally the Constitutional Convention as it happened, which was technically illegal according to the previous laws. But once you gathered a few hundred of the most prominent Americans who could really say no, right? It was impossible to say no, right? And there's this very fundamental way where the system has sort of run very well. It's clear that the United States is, you know, arguably like the greatest political system so far in human history, just in terms of its wealth, power, and influence. I'm not going to go into, you know, its many moral virtues and many moral failings, but it's surely, you know, one of the most powerful countries to have ever existed.

1:27:31And, you know, that is definitely the sort of thing that will put Americans in the history books next to the Romans, the Chinese, and so on. But to go back to sort of the great founder theory perspective, like the founding of religions, the founding of states, the founding of new types of economic organizations, right? the founding of new cities. You look at the history of it, often it's like very, very concentrated and they come into existence very rapidly. You sometimes have long periods of stasis followed by a very short period of development followed by a much slower rate of change. Presumably, and by short, you mean short enough for a single lifetime, which would suggest that it was basically one person or a small number of people working together to sort of design new rules of the game.

1:28:20Exactly, which is precisely what we saw with the American Revolution. It's precisely what we saw with, like, you know, the big religious transformations of history and big military or economic transformations of history. So having said all of this, I think that in the past, whenever you have a society trapped in negative-sum games, It is precisely on the creation of new social technologies where the hope rests for positive sum games. So I think today we are hopefully generating enough people who have the potential to be great founders that there will be a transformation in some sectors of our society.

1:29:03that gives us from a losing streak where we're sort of living off our inheritance, where maybe the numbers are still going up and maybe even the first derivative is positive, but the second derivative has gotten very negative. So we need to break out of that spiral. And I think that can only happen through such individuals. To push back on great founder theory, I wonder whether, because presumably with history, there's going to be these sort of blind spots and biases in the way that the information has been recorded like we love to think of things through stories right and how what usually happens in a story that there's a singular protagonist and we can go okay well they did that and they wanted this and therefore this happened stories aren't very good at telling and nor is written history telling the story like telling of situations where it's kind of a decentralized bottom-up type thing that emerges so i i worry whether it's like taking a kind of overly deterministic sort of view of of history or at least downplaying the effects of emergent things that have happened what would you say to that i would say that definitely there are many very important and emergent phenomena that have happened through the course of history.

1:30:23In fact, it's very useful to study those and very useful to try to model those great trends. Having said that, though, I think that precisely because of our limits of modeling a very complex reality, we have to focus on the most explanatory bits, right? And I think that while great founder theory is not supposed to be a total theory of all social phenomena, it does a rather good job of tracking the following, rise and fall of organized religions, rise and fall of political systems, the great technological or economic transformations of several eras. And you know, if you got those three, I think you're doing pretty well, right?

1:31:08And I think that so much of what we would consider bottom-up change, so much of it rests on changes in the macro environment. It's sort of like in a desert when the climate pattern shifts and the rain starts to fall more frequently. Of course, the seedlings growing everywhere, sprouting, and eventually the trees growing. Of course, that's a decentralized phenomenon. But it might be just a side effect of a single weather pattern changing and there being an extra monsoon there, right? So if you consider something like the introduction of a new code of law, of course, of course, the work, people working under incentives of the code of law are, that's decentralized, right?

1:31:58People suddenly have an incentive to open or create a small business. They create a small business that wasn't centrally planned. However, the transformation from, say, I don't know, a Soviet-style planned economy to an open competitive economy, that had to be a centralized change. People might argue that, oh, but the Soviet Union, you know, we can focus on that concrete example. Soviet Union fell for this or that inevitable reason. As far as we can tell, it was sort of just this coup d 'etat of Boris Yeltsin versus Gorbachev, where the constituent republics, the heads of those republics, who are already very powerful people in the Soviet Union sort of got together and they're like, do we even need Gorbachev?

1:32:39Do we even need the central system? Let's split this thing apart. And when they did, this meant liberation for many of the Baltic states. This meant a chance at independent development for Ukraine. But for many places, like the initial very powerful person that made that decision just stayed in charge as dictator, as happened in a lot of Central Asian republics. So history offers many, many natural experiments where we get to see, because consider, if it's truly the collapse of the Soviet Union was truly such a bottom-up phenomena of like freedom versus dictatorship, surely those Central Asian dictatorships should have been basically impossible.

1:33:15Like the same bottom-up that overthrew the Soviet Union should have overthrown them, right, in term. And maybe one could argue, oh, there was a delayed reaction. It was, you know, it had to wait until the so-called colored revolutions in the 2000s. But at that point, we're dealing with the effects of the internet. Society has changed so much. Look, that's 20 years after the fall of the Soviet Union. I think it's a very tenuous argument that that's what happened, right? So what do you suspect the common traits are that is shared by great founders? And what type of personality traits would be necessary for great founders or game designers, whatever you want to call them, that would be necessary to help society out of its current funk today?

1:34:00Well, those are two huge questions. And, you know, you've asked many huge questions today, which I struggle to answer. Doing very well. I strive to answer. Yeah, yeah. But to focus on maybe the first one, I think that nearly everyone has to thread the very delicate needle of being capable of disagreeing with nearly everyone in your society without needing to disagree with everyone. Is it not having the personal need of egoic attachment to it or what? Anyone who is always attached to disproving what everyone else is saying will never be able to work with other people well enough to transform even a single organization, let alone an industry or a society or a civilization.

1:34:46You have to be—so the people are sometimes somewhat megalomaniacal, but they can't be too detached from reality because with detachment from reality comes a lack of ability to survey the landscape, right? There has to be this sort of threaded needle where they, either for deluded reasons or good reasons, believe they personally have the moral authority to transform society. They might believe that they're chosen by God, or that all that matters is technological progress, and I'm making technological progress happen, or that I'm producing a wealthier society, and my opponents are sort of like the Randian villains of this story.

1:35:27something has to be there that morally justifies first to yourself and then to your close allies action because human defaults tend toward inaction and i think we are extremely good as a species of bullying each other out of doing things i think this is an under emphasized fact of human nature that we are very much uh you know we are very much a species that does not like the tall poppy we try to cut it down we're sort of fascinated in an almost salacious way when someone gets away with being exceptional we kind of worship them we over focus on them and when they slip up just a little we're vicious ready to pounce so our excitement is more kind of this uh it's very unflattering i think i think we can be quite abusive to the exceptional people in our societies and always have been as a species because the rewards are very high possibly of successfully being such a person, but the punishment is always there.

1:36:24The most dangerous profession you could have in the Roman Empire was being emperor. Your odds of a violent death were very, very high. And I think in the modern United States, the most dangerous profession you can have is actually still technically the president. Your odds of being assassinated are, what, 3 %? That's pretty dangerous. There are few jobs where I have a 3 % chance of being murdered. Maybe like a military event when the US is actually in a serious war. Maybe that comes close, but I don't even think that comes close. So that tells you something interesting, right? Even though we're much less violent than Rome ever was or Sumerian society ever was, we're a very peaceful society, it tells you that there's an exposure for violence, which leads me to feature number two.

1:37:08I think there's something about sheer tolerance for stress, sheer physiological tolerance for cortisol. That has to be much higher in that type of person. And then the third feature is the ability to have sustained long-term allies. And again, this is a place where a good story often is just true. If you examine independent or hostile accounts for discussing the life of, say, I don't know, Muhammad or of Julius Caesar, even their enemies acknowledge that they have close allies who follow them for most of their lives, who they can rely on, who they can confide secrets in, who they could even voice doubts to.

1:37:54Because no individual, no matter, again, how effectively megalomaniacal or how high their stress resilience is, can process all the information truly. All they can do is be sort of a clearinghouse for the cross-information of a trusted small group. And even the most caricatured historical figures, like let's look at Napoleon, right? He relied greatly on his military commanders and on his legal experts. And I wouldn't say that he was necessarily a great founder unless, say, French civilization endures for another 500 years, because the code of law that he promulgated, Napoleon's code of law, is actually the basis of civil law all over continental Europe.

1:38:38So if that system proves to be successful, we could retroactively say that Napoleon was a defeated military commander, but despite that, still a great founder. If he successfully introduced the code of law, his key ability was, despite his cult of personality, to still have close allies and people that he could work with for years or decades. I mean, it's clear that some people do have more of those traits. I know some of them personally. as you're going through that list I'm like yeah that fits the bill on a number of like particularly like powerful CEOs or whoever people who are um they're just incredibly resilient and they aren't afraid of like challenging opinions and so on but it does seem to be the case that certainly fame and to a large degree the power that comes with it or just power even without fame will eventually fry your brain.

1:39:31I have not yet seen anyone whose brain has not been at least partially fried by that. So how to manage that if we want to find people who are able to think through these, hold these incredibly complex issues in their heads sufficiently and design structures and have enough of a cult of personality to bring people who will help them bring these things to life, how to stop them from then getting drunk on their own success? Well, there are certainly ways to ameliorate that problem, but they may perhaps propose that in the big picture, this is less of a problem than we often fear. If through the course of revolutionizing an industry, of say through the course of Steve Jobs giving society the iPhone, right?

1:40:16Or through the course of Buddha launching a self-help movement that in the West we'll call a religion, let's say, or Charlemagne breaking the constant tribal conflicts and producing the system we think of as feudalism, they go a little bit insane. I mean, that sure isn't nice for them. It probably isn't nice for people towards the end of their career. But from the perspective of civilization as a whole, they've already delivered their payload. In biology, we're often like, well, after you reproduce, it sure is unpleasant where your body falls apart, but it's what natural selection has selected for.

1:40:58And unfortunately, maybe like, you know, it is the cost of this process of social creation, the creation of a new social technology that has to be, you can't just write about your utopia. You have to build a prototype community that actually carries out these ideas. And look, your prototype community could be, I don't know, Plato's Academy, sorry it could be Aristotle's Academy right it could be Plato's school it could be a new city state it could be a new religion once you've built it it is bigger than you it outlasts you and even if you do other things that are bad for society or bad for you or those close to you I think society has been vastly improved by your presence by your existence by that initial act of creation I could be a little bit provocative and you know point out that you know uh you know a scientific analogy here would be maybe Fritz Haber, who, you know, is very important for the Haber-Bosch process.

1:41:53It's how we create all our artificial fertilizers. This guy in the early 1900s, I think probably responsible for at least like half of the yields of the so-called green revolution, which allowed us to feed billions of people. So really saved many, many millions of lives. On the other hand, though, also, you know, worked for the government in World War One and made poison gas. And I'm sure that killed a few thousand or tens of thousands of people. And then on that, it sure would be nice if we could have a clean utilitarian hero that just did the farming revolution and didn't do the poison gas. But really, he was the kind of person that was up for either and was convinced that this was sort of both good, right?

1:42:36He believed and my side should win the war, not the other. Chemistry will revolutionize all aspects of life. And really that sort of positive and negative thing I think often goes together. So my argument would be the following. I think that as a society, we need stem cells. In the body of society, we need the stem cells that go around the body and become new skin, new hair, new nerve cells, right? Right. Highly adaptable. Right. And if we downregulate the amount of such stem cells, we will age and we will die as a society. If we upregulate it, yes, also the risk of cancer goes up. But on that, I think we really need renewal that badly.

1:43:24So I really do think that the sort of power issue, I think so much more evil in the world has happened through something not being done than through something being done. It's very lopsided. Even if we put Genghis Khan and all the terrible conquerors of history on the pile, and we put them next to all the city builders who are also often conquerors, to be honest. we just talked about, Alexander, raised many cities, built many cities. On that, probably good for humanity. Probably the library of Alexandria was worth it. But it's a complicated equation. But if we put those people on a pile, and then we add cultural reform, and scientific reform, and weird groups pursuing various aspects of knowledge, I think it clearly comes on the side of, on net, human agency is good.

1:44:14And in the bottom line, when we talk about power, we're talking about human agency, right? Maybe it is the case that humans can only endure so much agency, and with too much agency, too much optionality, our brain goes awry. But I think fundamentally, we are better off with a higher agency set point than a lower one. We are nowhere near the frontier of crazy, irresponsible people running around. We're very close to saturated on extremely boring, very parasitical, very inactive, extractive elites. So we're close to saturation on the senescent cells. We're nowhere near on saturation on new crazy people running around.

1:44:56You could argue, though, that we are now in a world where through technology, and particularly information technology, that such power is becoming increasingly democratized. so well I'm curious because like I just had Nick Bostrom on he you know his thesis on the vulnerable world hypothesis talks about um for the most part most technologies are great and that's good that we are democratizing them but sometimes you pull out a bad technology that is actually incredibly damaging and if you democratize that make it available to everyone then the individual the the ability for the individual to fuck everything up is has increased so So wouldn't that, you're saying that actually the rate of democratization of power is insufficient compared to the stasis?

1:45:45Well, the way I would propose it is that individuals or small groups wielding power, pursuing agentic agendas happens much less than we think it does, right? Right. Most elites, extremely passive, extremely homeostasis seeking. Right. Only protecting, you know, I use the phrase class interest or class privilege previously. They've got power and they just want to keep it. Exactly. And they want to give it to people like them. Right. And it's not explicitly thought of. It's just like, oh, I get along so well with this person. And, you know, we met in college. And which college was that? And we were in the same frat.

1:46:21And the frat was the skull and bones. right? It's very not self-aware. And people sometimes hallucinate and imagine that there are agentic elites elsewhere. And then you're like, but wait, I've just talked to all the top academics of this country. I've just talked to several people who are, you know, members, members of parliament, and I've talked to the most prominent businessmen. And in a small country, like maybe Slovenia, or maybe a little bit bigger like Britain, where are these agentic movers and shakers. Everyone assumes that someone else is out there improving society. But really, if you're in an elite position and your social circle is in an elite position and you see most of them are not doing much, that means no one in your society is doing much.

1:47:04Okay. With that point sort of out of the way, to address what I think is sort of there in your question with regard to democratization, I am skeptical of the democratization of our society. For a variety of reasons I think our society is very oligarchic and power has actually concentrated in these sort of like ossified, stagnant, non-agentic centers where the main mode of exercising power today, say in the European Union or in the United States, and honestly any large country, is the veto. In the European Union, the most powerful person is a French-speaking bureaucrat in Brussels that you've never heard of who says, no.

1:47:49Says no to something being done. And in the United States, there are so many reasons not to do things, right, at the federal level, at the sort of political level. And we are saturated in these various elites. They are not very agentic. They are powerful. But because they don't do anything, they just forbid things, you can never call gotcha on the power. We don't recognize forbidding something as an abuse of power. We recognize doing something as an abuse of power. In case it goes wrong, yeah. And that's kind of messed up. Definitely. It's an asymmetry of information, essentially. Very fundamentally.

1:48:25And then we were just talking about the failure of the epistemic commons. I think there is no way that power is more democratized now than it was 20 years ago because, you know, my phone is not my ally in making sense of the world. It is actively trying to run basically psyops on me at all times and all places. and you know by carrying around this phone or an apple phone if people prefer those you are basically giving away your location at all times you are basically recording every conversation you have and while you might trust the very inert and uncreative elites that run the big software companies or that run the government agencies that can legally demand any data your big software organizations have, your companies, your nonprofit foundations, whatever, while you might trust their inactivity, the fact is you as an individual are vastly more transparent to any bureaucracy today than you were at any previous point in human history.

1:49:25In no previous point in human history could a bureaucracy gather as much information on you as an individual as they can today. And I claim that you as an individual have less of an ability to gather information on the inner working of a bureaucracy today than even 10 years ago. I think even compliance with things like the Freedom of Information Act, right, have been much attenuated. And the idea that it is irresponsible for large bureaucracies to not keep accurate records, that's sort of gone away. And the software and information tools that an individual could use to study those organizations, they're not been developed.

1:50:04We've only developed ways to monetize and examine the individual. We've not developed the software tools to look at large bureaucracies. Ergo, over time, large bureaucracies are more powerful than individuals and large bureaucracies, unless piloted by a live player, so like an active agentic individual, will come to be run by homeostasis-seeking, self-serving, rent-extracting oligarchies. So it sounds really like what you're arguing for is we just need actually a higher complexity state. We need greater risk tolerance. I would say we need greater risk tolerance for reform of major government institutions.

1:50:41And we need greater risk tolerance in terms of industry disruption. And we need greater intellectual risk tolerance where we encourage people to, and we try to sustain them socioeconomically in developing new ideas of how nature works and how society works. What advice would you give to anyone watching who feels like they resonate with this idea of a great founder and want to try and be one themselves? I mean, first off, it's very excellent to try to be one yourself. I just selfishly encourage you because we all benefit from the externalities on net. But secondly, I also propose perhaps finding a great founder and trying to be part of their team and trying to be a functional member of a small group might be a more practical first step.

1:51:32where I think in our society, we've lost a lot of this idea of recognizing and valuing greatness in others. There's perhaps a way in which I might be slightly less individualist, like a lot of my philosophy can be described as this very hyper-individualist, maybe almost Nietzschean will to power. But one of the things I also believe is in valuing, appreciating, and supporting human excellence. I think there's an undersupply of people seeking out the most agentic people in society and attempting to enhance them and strengthen them and help them correct their flaws and assist them maybe with going a little bit less crazy or being a trusted team member and long-term associate.

1:52:16A part of lost Western social technology is, I think, actually the loss of a load-bearing type of friendship that once existed, right? Whether it be the sort of like scientific legiality where you both care about the truth and you're actually collaboratively working towards it, whether it's the kind of, you know, board when the first, you know, when the East India Company met to do, you know, their business or whatever. A very trivial thing is that they would open the meetings with prayer because the pretense was that actually what we're trying to do here is work each other's salvation, and we're just trying to do good work together.

1:52:55And only incidentally are we super rich and making a lot of money and being super powerful. And I think the idea of valuing other individuals so much that you want to help them fully develop their potential, that is something that was there in the DNA of Western civilization. It was there even in classical civilization. If you read Plato and Aristotle on friendship, they elevated to this level that we just don't understand, it's elevated even above romantic love in the writing, right? I think that allowed very functional small groups to work together. So I would propose, find someone in your life who you think has great potential and do your best to mentor them or assist them or learn from them or teach them to try to correct their flaws and be constructive, right?

1:53:46But don't say no, say yes to things of others. And I think the adage of, you know, there's no limit to the good you can do if you don't care who gets the credit will prove very true. All right. So to finish up, I would love to ask you some rapid fire questions around geopolitical predictions. Because you are an analyst. ultimately one of your main things you do is you run Bismarck Analysis, which is essentially an intelligence gathering resource. Basically private intelligence in a research and consulting firm. We also put out the weekly Bismarck Brief, which is well-received and has a very interesting set of subscribers.

1:54:32I certainly practice what I preach, right? I try to give accurate evaluations of the world. So fire away. Okay. Percentage likelihood that TikTok gets removed from the App Store in the US by the end of 2025? I think that the US regulatory apparatus moves slowly and the countermeasures are fairly effective. So I would say that about 20%. Likelihood that Ukraine joins NATO by end of 2030? Okay, I would put that at about 35%. Likelihood that COVID was a lab leak? 75%, 80%, 80%, yeah. Likelihood Netanyahu stands down by end of 2024? I don't have an excellent model of Israeli domestic politics. I think a lot of the details there are opaque, mostly because of the very competent security services of the country.

1:55:33I would bet that it's likelier than not that Netanyahu endures just because he is such an effective politician, but with low confidence. Likelihood that China invades Taiwan by end of 2026? Fairly low. I would put it under 10%. Finally, likelihood that Epstein killed himself? 15%. Okay, it's pretty low. Awesome. Any other big predictions you'd like to make that you are very high confidence on? Okay, well, that's an interesting challenge in its own right. I predict in three years, or actually, no, I predict in five years, most of us will still have a job, and AI will not have fired us. Maybe it'll make our job much more boring, kind of lame and rote.

1:56:28But you know, I have a quip that says, you know, it's impossible to automate fake jobs. I think our society has a high degree of jobs that only exist because of sort of regulatory and social needs. And that actually we have secretly already implemented UBI. It's called having a middle class job. And it's just sort of janky through all of these like hyper regulated types of ways in which we redistribute wealth unofficially. So you think that actually a lot of people going to an office is purely performative? And I think it doesn't mean that they're having a good time. I'm just saying that our society has a political need for jobs that exceeds its economic need.

1:57:07And while AI advancements might well come to the point where they suppress the economic need or meet the economic need, they do nothing for the political need for jobs. I do think it's, for example, telling them when it comes to large tech companies, it's extremely unclear what the median software engineer does. I think you could easily go through every single major tech company and fire half of the software engineers and you'd be okay. And the user wouldn't notice a big difference. I mean, Elon did it with Twitter, right? Elon did. 75%, yeah. Yeah, yeah, yeah. But, you know, he had to know what he was doing for that.

1:57:44But I almost think that if you didn't know what you were doing and just magic eight balled it and fired half of them, the other half would scramble to do that job pretty well. And if that's true in tech, which is actually a very demanding field, very rigorous, technically demanding, what does that say about so many other sectors of our society? I think we are inundated with make work that secretly is a type of work. It's just political work. We don't think of it as political work. But it kind of is, right? Like what's the point of having a large team of engineers if it makes the product worse?

1:58:19Well, it means that the manager of the engineering team claims greater authority, can claim like more credit, might advance inside the hierarchy. What's the point of a big tech company having 10 ,000 employees rather than 1 ,000 employees? It's sort of like, you know, Mr. Congressman, we bring 10 ,000 high-paying jobs to California. So it's a signaling thing again. Exactly. And why do the elected officials care about whether there are 10 ,000 high-paying jobs in Austin? Well, that answers itself, doesn't it? So it sounds like actually perhaps the most promising disruption could come through lifting the veil on the signaling game.

1:58:55An effective accounting of the economy and politics would be a revolutionary act. And I think there's more effort that can be put to that. Thank you so much. Thank you. So there we go, everyone. Thank you so much for tuning in and huge thanks to Sammo for this conversation. In the show notes below are links to a lot of his writing. I really recommend taking a peruse. Also, if you want to keep your finger on the geopolitical pulse, do check out his weekly Bismarck brief. You can also give him a follow on Twitter. Lastly, quick request. Please, if you enjoyed this, give this podcast a good review on whatever platform you're listening on.

1:59:33I hate to ask, but it really helps. So anyway, thank you and I will catch you next week.