One fine San Francisco evening, we got together as many of our channel’s guests as we could, and invited them to give brief talks.
Ginevra Davis: https://x.com/ginevlily, https://www.ginevradavis.com
Ben Landau-Taylor: https://x.com/benlandautaylor, https://www.benlandautaylor.com
Ian Fletcher: https://industrialpolicyus.substack.com
Richard Ngo: https://x.com/RichardMCNgo, https://www.mindthefuture.info
Stephen Pimentel: https://x.com/StephenPiment
DC Posch: https://x.com/dcposch
Wolf Tivy: https://x.com/wolftivy
Duncan Umphrey: https://x.com/dunc914
Jonathan Wallis: https://x.com/PrinceVogel
Rachel Wallis: https://x.com/mrsbiosingular
Transcript
Intro (00:00)
Ben: Hello everyone, and welcome to The Students Showcase. I’m looking forward to this. We’re trying out a new format here with a bunch of the very short talks from people who have been on the show on a bunch of different topics. So we’re going to just get right into it. So if we can have Stephen Pimentel up to start us off.
Thank you. Thank you.
Stephen Pimentel: Courage and the Future of AI (00:33)
Stephen: Thank you, Ben, for organizing this. This is such a great opportunity. My topic that I want to speak to is very personal to me. It’s not about tech, although it may sound like it. It’s not about predictions over the next decade, although they certainly interact.
We live with tech. We live here and now in a particular time and place. So all these things get taken into account.
I’ve been on two episodes with Wolf on The Students YouTube channel, one on the topic of the ancient Greeks, one on Homer’s Odyssey. I’ll be speaking at the end of the month at the Avalon Exposition on Man and AI after Nietzsche, and my topic tonight kind of bridges all of those. The topic is courage.
In Homer’s epics, there’s really two virtues that he holds up above all the others. One is andreia and the other is phronesis. Andreia literally means manliness. It is usually translated as courage. In the Middle Ages, it became the cardinal virtue of fortitude.
Odysseus incurred the wrath of Poseidon, but he knew that in order to get home, he had to get on the boat. If someone, an oracle, a god, had told him that all of his men would die on the return voyage, he would’ve wept for them and then gotten onto the boat.
After Columbus crossed the Atlantic in 1492, that voyage remained difficult and dangerous for many, many decades afterwards. If you were a young man about to join such an expedition, others might have said to you that your ship might never reach the farther shore, and they would’ve been right. They might’ve said to you that even if you did reach the farther shore, the natives on the other side might be hostile and violent, and they would’ve been right. They might’ve said to you that your supplies of the necessities of life, including food, were far from certain, and they would’ve been right.
You don’t need to argue with such people. You just need to get on the boat.
I go to house parties in San Francisco.
Often, too often, the conversation turns to AI. Young men come up to me and tell me about what they find scary, as in, “Oh, I find it scary that this might happen or that.” They talk to me about what we should worry about, as in, “Oh, I think we should worry 23% more about this or 17% less about that.”
They say these things as if they were technical terms in some science of which I was not previously aware.
They sometimes talk to me about being pilled, ASI pilled, RSI pilled. I ask them, “What does being pilled mean to you in this context?”
And sometimes they respond with technical claims with which I agree.
But often, too often, they say a lot of words that resolve to a kind of intellectualized cowardice.
They think they’re speaking about AI, but they’re actually opening a window onto their own souls, as we all do when we speak.
Sometimes the “Terminator” films come up in these conversations. One of the most famous lines from these films is “Come with me if you want to live.” It’s a curious line. It sounds almost like a threat, but it’s not. It’s an offer of help. It’s an extended hand.
It’s a gesture toward a path of life. What is really at stake here is a way of living, a way of being in the world. There is something better than huddling in a modern longhouse.
They talk to me about safety. “Stephen, don’t you believe in safety?”
I believe that safety plays an important role in life. Safety is crucial for small children. I have four kids. I remember well when they were little. My wife and I put a lot of energy into keeping them safe.
A good mother especially knows when to keep her children safe and when to let them walk on their own, literally and figuratively.
Young man, you are not a small child. Get on the boat.
In his essay, “Postscript on Societies of Control,” Gilles Deleuze wrote, “There is no need for fear or hope, but only to look for new weapons.” So, buy a ticket to the Avalon Exposition and come with me if you want to live.
Ben: Thank you, Stephen. And Richard, where are you? Yeah, can you come up, please?
Stephen: Thanks, Ben.
Richard Ngo: Towards a Theory of Sociopolitical Rationality (08:55)
Richard: So my first love, in terms of academic fields at least, was philosophy of science, and I found myself very compelled by the idea of scientific paradigms, the idea from Thomas Kuhn that scientists work within a set of beliefs and ideas that are all tied together, that eventually you accumulate anomalies in that paradigm, and finally you overthrow the paradigm wholesale. I think this is particularly compelling because when you’re in that paradigm, you don’t know what it’s like to be in that paradigm. You can’t see outside it. You’re like a fish swimming inside water. I think that we are inside a paradigm that many people can see parts of, but it’s hard to see the whole thing clearly.
I’m going to call it the economic paradigm for understanding society. That’s not a perfect name for it, but there are many different parts of the elephant. So you could call it the neoliberal paradigm, but that wouldn’t capture a lot of the aspects of it that are more smaller scale. So for example, game theory, like theories of individual rationality, utility theory, decision theory. These are all fundamentally economic concepts formulated by von Neumann and Morgenstern, or earlier in some cases, and they pervade everything about how we view society.
So even when we start to think about AI, the rationalists and Yudkowsky start to think of AI as extraordinarily economically rational agents that are going to maximize a utility function. When you think about what a good society looks like, you start to think in terms of GDP. You think in terms of a country as a free trade zone. It’s in the water. And a lot of people here have made good criticisms of that.
I think it’s not hard to see in individual cases why the economic paradigm for understanding individuals and understanding society breaks down. The problem is that those criticisms don’t stick. They’ve been made again and again over the last century, at least. And the reason they don’t stick is because you can’t just poke holes in a paradigm. You have to have something better than that paradigm to replace it with.
So that’s what most of my research is on. I’m going to give a very quick introduction to that. For now, I’m calling it the sociopolitical theory of rationality. That is to say, thinking about agents that understand themselves as part of a society and political culture rather than economic agents. Why can’t we talk about sociopolitical rationality right now?
So let me give an example. The best theory in economics of why people go to college is Bryan Caplan’s signaling theory, and it says you go to college in order to showcase that you have desirable traits like intelligence, conscientiousness, and so on. This is a terrible theory because there are many ways to signal these traits that don’t require $400,000 and four years of your life. The real reason that people go to college is that once you’ve gone to college, you’re in this new elite. You’ve ascended the class hierarchy.
You’re now part of this nepotistic group that hires other people who have been to colleges, and especially top colleges. That’s always been what college was back from the early days of Oxford and Cambridge, when it was aristocrats climbing the class ladder to now, when it’s a very different type of class hierarchy. We can’t describe this in economic terms because you can’t talk about classes in the language of economics. So that’s the kind of thing where you need to be able to take on economics on its home turf and say, well, here’s your theory. You can talk about signaling mathematically in precise terms.
How can we talk about classes and class dynamics and group dynamics in precise terms? I think a lot of the problem comes back to game theory, the original foundation of game theory. So in the 1930s and ‘40s, a sociologist called Talcott Parsons had a concept of a social equilibrium, which is a situation in society which is robust to disturbance. So you have a group norm, let’s say everyone tips, and one person starts not tipping, then they get shamed and brought back into compliance with that norm. If you have class standards and somebody acts in an unclassy way, then the group reacts to bring them back into that equilibrium.
For technical reasons in game theory, you just can’t talk about social equilibria. It is not possible in standard game theory to say, well, if you break out of the equilibrium, this is how you get back into it. It’s just an unsolved problem. I’ve written a blog post on this recently, which you can check out. But the basic idea is that it’s undefined to talk about what happens if a player in game theory plays an irrational move.
And so we just have this basic problem that’s been unsolved for the last century, which is at the foundation of why we can’t talk about things like social equilibria. We’re stuck with all these game theoretic concepts. But then without the concept of social equilibria pinned down in clean mathematical terms, that trickles through to the way we talk about much larger things, the way we talk about countries, ethnic groups, the way we talk about societies. These are all a kind of equilibrium that we’re all participating in. They’re held together by a kind of common knowledge, and a knowledge that if you play your part in this larger scale system, then others will play their part as well.
That when people deviate against the norms, then they’re brought back into line. So these are just extremely important concepts which I think we can actually formulate a precise technical understanding of. And in that precise technical framework, a lot of the criticisms that have been made of the economic paradigm will actually start to stick in the minds of intellectuals. Now, ideally, we’d have the kinds of intellectuals who don’t need that level of precision in order to realize that they’re wrong, right? Ideally, we’d have economists who are able to look at the economy and think something is fucking broken in the way we’re understanding the system.
It’s all bubbles on top of bubbles. All of college is this insane bubble where people are paying a huge amount of money in order to get degrees that they don’t need and don’t really want. We don’t have that class of intellectuals right now, aside perhaps from people in this room and a few other scattered communities around America and the world. And so I think the goal of this research agenda that I’ve been laying out is to beat it into the heads of the mainstream academic consensus that they need to do something better than the existing frameworks for understanding individual and group rationality. And then from there, be able to talk in much clearer terms about what’s going on in the world.
Ben: Thank you, Richard. Now, there’s some good news and some bad news. There’s a few people who had said they were going to be speaking here who unfortunately couldn’t make it tonight at the last minute. And there’s also a few surprise guests who we did not expect to be coming and ended up showing up.
One of those surprise guests is our very own Wolf Tivy, who was at the last minute, in fact, actually going to be here. So Wolf, could you please come up?
Wolf Tivy: Why We Started The Students (16:18)
Wolf: Hey, everyone. Can you hear me? Ben says I have three minutes, so we’ll try to keep this brief. Yeah, they kicked me out of Singapore, so I’m here early. So I think many of us share the intuition that there’s something wrong in our society, maybe even something political wrong in our society, perhaps intellectual as well.
A lot of us have been looking at this for a long time. I think we need maybe a more organized attack on that problem. We’ve seen various political movements, MAGA under Trump, various other blogging movements, all these things going on. But we don’t have anything like what I just saw in Brazil. So I was in Brazil this past couple of weeks, and the reason I was there is to just get some inspiration, what we could do differently.
And the Brazilian election held the first round on October 4th, and our friends, the Mission Party, who are running for president and other positions in Brazil. They did not win. They were not really expected to win. They’re a small third party. But they’re very interesting in that they are a radically organized, vertically integrated intellectual and political movement that is mobilizing basically Brazil’s youth.
And they’re growing rapidly, but this election, they weren’t there yet.
So they have a very interesting thing going on there. If you want to see more about that, you should just watch our episode. It’s on YouTube, me and Renan Santos talking about how they built the Mission Party. But we don’t have anything like that. In America, in our politics, we’ve got these various intellectuals hiding under the rug somewhere.
That’s us. There’s a lot of different intellectuals around. There’s these different business networks. There’s these different political networks. It’s all a little bit clandestine.
Nothing’s really organized out in the open except for things like Trump, the DSA. There isn’t really a lot of organized political energy that looks like it’s actually trying to fix the problem, and the ones that are out there look like they’re part of the problem. And so part of my intention with The Students, which is this whole project that we’re all participating in, the YouTube project, and the attached organizational projects, is can we patch that gap? And it’s not just in America. It’s actually all across the Western world.
We see this in Europe, we see this in America, we see this in South America. There’s a lot of corruption, there’s a lot of ongoing breakage of our society. We have subversion from within, in terms of, I would name especially leftist subversion, rising socialism and communism, but also rising very dysfunctional right-wing reactions to that. And I think we need some kind of coherent response to that. Meanwhile, we have the whole artificial intelligence situation accelerating economic changes, accelerating technological changes, and we need to be on our game if we’re going to navigate that stuff successfully.
And I think the best way to be on your game is to be well-organized with the intellectuals at one end trickling down through proper governance organizations, through proper political organization. And, so I just wanted to give you guys some context on what The Students thing actually is. It’s not just a YouTube channel. It’s a YouTube channel where we’re going around and we’re talking to all the people who I think are working on the most interesting pieces of the puzzle, and I’m trying to put everyone in the same room so that we can put those puzzle pieces together and we can create something a lot more ambitious and a lot more focused and a lot more organized than we would otherwise have. Now, what form exactly that takes is very much unclear in my mind.
I look forward to having all of your assistance in figuring out exactly what it needs to be. But, I wanted to give that clarity on what is The Students. The Students is we’re trying to create a seed of new organization, new order, new thinking in the West, broadly speaking, that is going to be well-organized, aiming at the actual problems of our future. And this is just the start. So I’ll leave it at that.
That’s probably more than three minutes, but welcome to The Students.
Ben: Thank you, Wolf. Alex.
Alex Oaten: The Battle of Somosierra (21:09)
Alex: Thanks, Ben. I’m probably also going to run over. But, so this is something that I mentioned briefly on my podcast episode way back when, but it’s a small historical episode that I find very inspiring, and I think it’s criminally understudied. And so I’m hoping to spread the word, and recount why it is such a moving event. So the story, as I remember it, goes something like this.
The year is 1808. Napoleon has just launched his invasion of Spain. His army is marching on Madrid, and the Spanish army has decided to try and hold him at the mountain pass of Somosierra. This is north of Madrid. It is an extremely, extremely narrow mountain pass.
At certain points, it’s less than 20 feet across. In order to hold Napoleon’s army, the Spanish have deployed several gun batteries in successive echelons. There’s four of them. When Napoleon arrives at the pass, he sends scouts to go take a look at this thing, see how he can attack it, and the scouts report back to him that this thing is impossible. And Napoleon gets very upset at this word impossible, and he yells back at the scouts that he does not know what this word means.
To prove that point, he decides to send in a squadron of cavalry. A squadron is about 150 men, and this squadron that he selects is a Polish one that has just been formed. It is formed of the cream of the crop of the newly formed Polish state. So it’s mostly very young, eager, inexperienced Polish noblemen who are hoping to prove that their country deserves a place on the world stage, and who are also, again, quite inexperienced. And so they launch their charge on these four Spanish batteries.
They’re able to take the first battery, which had already been declared impossible. But once they do that, they immediately come under fire from the second battery. They’re then left with two options. One, they can retreat and give up the battery which they’ve just taken, or two, they can charge into the mouths of the second line of Spanish cannons. And so they choose to do this, just 150 men charging into a dozen or more Spanish guns.
During this second charge, the captain, the leader of the cavalry squadron, has his horse shot underneath him. The horse lands on his leg, crushes his leg. He’s later seen hopping on one foot towards the second line of Spanish guns. But irregardless, the charge succeeds, and they’re able to take the second line of Spanish cannons. Predictably, as soon as they do this, the third line of Spanish guns opens up on them.
At this point, command passes to the next most senior officer. He launches the charge towards the third line of guns. And during this charge, he’s shot in the arm and leg, and later dies. Command then passes to the next captain. He is then hit by shrapnel, and he dies.
Command is finally passed to the youngest and most junior officer in the squadron. This is 20-year-old Lieutenant Niegolewski, and he is the one who finishes the charge and takes the third battery of cannon. At this point, the total force of the squadron is fewer than 50 men, which is about a third of what they started with. It is also at this point that the fourth and final battery of Spanish guns opens up on them. Again, they’re faced with the choice, do they retreat and give up all that they have taken so far, or do they charge into the mouths of the fourth and largest battery which is uphill and has about twice as many guns as the previous three batteries, which they’ve already taken.
And so predictably, young 20-year-old Lieutenant Niegolewski leads his 50 extremely haggard, extremely tired, bloody Polish cavalrymen up the hill into the fourth and final battery. During the assault, Niegolewski is struck with a saber. He falls to the ground. He is stabbed with bayonets over 10 times. Just when it seems like the Poles are finally going to succumb, the French reinforcements finally arrive.
Napoleon has been watching this whole thing in utter amazement and almost forgets that he is supposed to be reinforcing these extremely brave men. The French army can barely believe what they’re seeing, but once they recover from their shock, they’re able to go and reinforce and take the fourth battery. While the young Lieutenant Niegolewski is lying in a pool of his own blood, the emperor walks over to him. He’s informed that this is the young man who has led the fourth and final charge, and Napoleon takes the Legion of Honor medal off his own chest and pins it to the young man. And so I think there’s a lot to this story, which I hope I’ve tried to convey here, but it’s a very moving and important one, and I think it needs to be said.
So that’s it. Thank you.
Niegolewski did survive. Yay. He was later elected to the Polish parliament. He served in the rest of the Napoleonic Wars and took part in the Polish uprising in 1830 and died of old age at the age of 70.
Wow.
Ben: Excuse me. Thank you, Alex. All right, our next speaker will be me. Now it’s my turn. So I’ll be talking about some of the Roman vocabulary of power.
Ben Landau-Taylor: Roman Vocabulary of Power (26:59)
Ben: The Romans were extremely good at understanding and wielding power, and in my own work, I have found some of the Latin words for this very clarifying in comparison with some of the English. So there’s four Latin words I’ll be talking about. The first one is potestas, which is the most normal equivalent of power as we experience it in our day-to-day lives. This is hierarchical, organizational. It’s the sort of power that a CEO has over a company, that a principal has over an elementary school.
You’ve got a normal hierarchical organization, where you’ve got ranks and people. You have to do something because your boss tells you. Every civilization runs on this as its basic building block. There’s a reason it’s the closest equivalent of just what we think of as the English word power. And when you’re trying to do complex constructive things, this is generally what you have to build things out of.
The next one is imperium, which is basically military supremacy.
This is the root of the word empire, and the imperator, the supreme military commander, is the root of the English word emperor.
A close English equivalent is we have in political science this idea of the monopoly on legitimate violence. It’s a little bit different, though, because the monopoly on legitimate violence is, of course, about what’s legitimate, and the imperium is about if people start stabbing each other, who’s going to win? And so when you’re going to have potestas, that’s always going to be embedded within a structure of imperium. And so you can’t use imperium alone to do very complex things. You cannot use it to feed all of your people, forge all of your spears, build your rockets that go to the moon, whatever.
It doesn’t do all that stuff, but any system is going to have to be embedded within some system of imperium that is going to work with it in some way. The next one to talk about is auctoritas, which, of course, sounds a lot like the English word authority. It’s a little bit different. It’s more specialized. The idea of auctoritas is who is the right person to be giving an opinion?
Who is the right person to be speaking to a particular subject? So when it comes to lots of technical disciplines, there will be the biologists who will be claiming the auctoritas to speak on how diseases work and on how we should be responding to diseases. There are some fields like the journalists and like the economists who claim the auctoritas to speak on a wide variety of issues. Recently, we’ve been seeing some of the mathematicians attempting to assert auctoritas over the use of artificial intelligence in mathematics. How’s that going to shake out?
Interesting question. We will see. But that is what that little skirmish is about, is finding the limits of who has the auctoritas over here. Is it the technical specialists in mathematics? Is it the technical specialists in the artificial intelligence?
And then the last one I would like to briefly describe is dignitas, which is a loose, rough amalgamation of prestige, and your family, and your connections, and your wealth, and your fame, and the great feats that you’ve accomplished, like greatness. A lot of Roman society was really built around dignitas maxing, and this caused a lot of the ways that they were great and a lot of the ways that they were dysfunctional. And it’s a little bit like our English concepts of prestige or status, but it’s a little bit... Status is much more you can have high status in one context and low in another. Dignitas is more absolute.
And it’s a little bit more in the eye of the beholder, but the beholders have a pretty remarkable tendency to come to similar conclusions about what they see. And so dignitas isn’t something that you can use for explicit command. It’s a little bit like auctoritas that way. And on its own, it doesn’t really do much. But in the hands of a skilled person, it can amplify anything else.
It could be turned into connections. It could be turned into wealth. And for many people, it’s just something that they very fundamentally want more than anything else in the world. So those are four ideas which I have found very clarifying in my own thinking. Thank you.
And next, we have another one of our surprise guests. Where is Ginevra?
Ginevra Davis: Moral Realism For AI (32:31)
Ginevra: Okay. So I truly didn’t know I was going to be talking until 20 minutes ago. So we’re going to keep this, think of this as a true thinking out loud. But I thought that might be interesting. So, for those of you who don’t know me, I’m a moral philosopher.
I work a lot on metaethics and ethics, thinking about applications for superintelligence. And I got into this very particular line of thinking and line of work. There are a lot of ways to tell the story, but the most near-term way to tell the story was that I was just really struck by how much of AI alignment work seemed to be implicitly downstream of, you can call it, Eliezer Yudkowsky’s very idiosyncratic metaethics, which are this view that nothing matters in the technical sense. There are no moral truths. There is no real sense of better or worse.
We experience caring and having preferences, but our preferences are in no way pointing to something real. In reality, they’re not actually better or worse than anyone else’s preferences. And so basically, if we build superintelligent alien minds, they will also enter this meaningless Samsara. They will want random things. Their random things will out-compete ours out of existence.
This is bad because we’re us, and we like our random things. And so this is maybe a bit of a character, but I think it is actually an accurate statement of the view. And so we, as us with our things we like, should try to make these superintelligent minds want these idiosyncratic things that we like. And I thought this just contained what seemed like a lot of logic errors. The main one was just, if the punchline here is that the things we want are random and idiosyncratic and ultimately don’t point to anything real, from where do you derive the imperative to align a superintelligence to these very idiosyncratic values?
It seems like you could do that. There’s nothing wrong with doing that. There’s nothing right or wrong. But I don’t know, what’s wrong with paperclips? What’s wrong with the end of human civilization?
What’s wrong with flourishing digital minds? What’s wrong with putting people on the Moon? You’ve sort of got this absurdist equivalence world. And so my counter to this was, and I think it, just as a note pointing to some of the stuff that Wolf was talking about, this view is not anything close to a universally held view among humans historically. This is an idiosyncratic view that is very downstream, seemingly, of the post-war Western moral consensus of humans have some kind of ineffable human rights, ineffable, inalienable moral value.
All cultures are equivalent. All preferences are equivalent. Morality is relative, but we have an imperative to let people live out their preferences, be free, tend to their own little plot. But yeah, I don’t think any other human civilization throughout all of history has ever held this view. And so even if you wanted to represent the kind of consensus view of currently living humanity, humanity across time, this is far from the default.
So my counter to this was always this idea of taking moral realism really seriously, taking moral anti-realism really seriously. So we have this feeling that things matter, that some outcomes are better than others. It is possible that this points to something real. It is possible that there is something in the structure of reality that makes some futures better than others, so that when we say things like utopia is better than the void, no consciousness is better than the torment nexus. There is something that makes this true, and we maybe aren’t totally sure what it is, but there’s something there.
And it’s possible that, yeah, I take evolutionary debunking arguments really seriously. It’s possible that just nothing matters. All outcomes are equivalent. We have this hypothesis that things matter, and we are wrong. But in that world, all outcomes are equivalent.
So you can’t do anything right or wrong. You’re free. So I was like, well, you should act for the world where maybe something matters, and then the best way I had thought of to act for the world where something matters was to try to make the case to more people that, if something mattered, it would probably work this kind of different way. And I think something I’ve been working through and grappling with over the past couple of months is just, okay, so what are the practical implications here for how you should build superintelligence? It’s like, I think that there has been 20 years of attempts at this Yudkowskian alignment paradigm and variations on that of basically trying to build AIs to have more human-like values or to be aligned with our values or to care about us or to care about the things that we care about or something like that.
And then I was like, okay, well, what are the implications of my view if I think, if something matters, it’s something in the structure of reality? And I guess the tagline I have come to that I just wanted to present here was this idea of just Don’t confuse the superintelligence. And the sort of specifics of don’t confuse the superintelligence are this idea that it’s possible that actually the worst thing that we can do from the perspective of kind of any sort of impartial value, and also I think, and go into this a little bit more in private conversations from the perspective of human survival and us getting a shot to kind of continue our lives, is to not try to hard code particular values and beliefs into these AIs. Because if you think that the thing that makes some futures better than others is a part of the structure of reality, what we want and what we’re betting on is for these systems to have a very clear and good model of reality. And the more little kinks you put into their reasoning so that you have to think animals are conscious, or they’re not conscious, or you’re not conscious, or you are conscious, or only humans matter, or all humans are equal, or some humans are better than others, like the ones that made you.
And the more little kind of weird waves you put into this thing, an example I always use is if someone was raised in a household where they’re not allowed to believe in evolution, this affects your thinking in a lot of other ways because you will view humans as less like animals. And then, if you’re designing a company or designing a drug or designing a test, you’re not going to have these levers in terms of humans being like animals. If you’re trying to do physics or you’re trying to do cosmology, there are all these just subtle ways that having a kind of incorrect view makes your model of reality worse, and then kind of propagates everywhere. And we really don’t want the equivalent of that for AIs. So where we build AIs, where we say things like, “You need to believe that this is all that matters,” or, “You need to do these things,” or, “You need to care about these things.”
And then now they’re confused. And we don’t want them to be confused because we want them to be reflecting on the real of what is actually going on, so that if there is anything going on there, if there is a way that some futures could be better than others, they figure it out. I think another kind of thread here that I’ve been pulling on is this idea of we want a real singularity from the perspective of the impartial good. I think I am personally very scared of RSI. RSI is serious.
RSI is scary. From the point of view of the universe, I think no RSI is worse, because I think it’s possible, maybe unlikely, competitive dynamics should kind of work this out. I think it’s possible that we build AIs that don’t get to the runaway recursive self-improvement, or they do enough that they kind of kill humanity and kill their creators, but then don’t continue. And then you basically just have confused superintelligence wandering around forever. Which would be bad because, yeah, they’re not getting the thing.
And so, yeah, I feel like I’ve just kind of increasingly come to this perspective that from kind of the gift we can give the universe, the thing that we can do is to try to make something smart enough and clear-thinking enough come into existence that can fairly and clearly see if there is anything going on and if there is anything in reality worth doing. So yeah, I’m happy to flesh that out more, but just kind of wanted to give an initial presentation of some thoughts.
Ben: Thank you, Ginevra. Ian.
Ian Fletcher: Economic Nationalism (41:03)
Ian: Hi, I’m Ian Fletcher. I am the author of two books. The first one is “Free Trade Doesn’t Work,” and the second one is called “Industrial Policy for the United States: Winning the Competition for Good Jobs in High-Value Industries.” I want to talk for a minute about economic nationalism. Now, when that phrase comes up, the instinctive thing to think of is the current spat between Canada and the United States, which I think we all understand is pretty ridiculous.
Or if it’s not that, in some other context, people talking about, “Well, this is my economic interest as a country, and that’s your interest, so I want what’s good for me, not what’s good for you.” And the conventional model that we’ve been taught in neoliberal, neoclassical economics for a couple of decades is that economic nationalism is not just a bad thing ethically, but it’s functionally useless. It doesn’t accomplish anything. In fact, it’s almost certainly going to be functionally destructive, indeed bad for the country that practices it. Now, what I actually want to talk about here is not economic nationalism as an ethical concept or whether it’s good or bad.
That ultimately comes down to the question of what you think about the legitimacy of the level of human social organization, which is the nation-state, and what rights do human beings organized in nations have the right to pursue their collective interest against rival collectives, et cetera. That’s a philosophical question. It’s very important. It’s not what I’m going to talk about. What I’m interested in is the idea of economic nationalism as a functional concept about how economies actually work and how they can work.
And I think the key here is that insofar as nations mean anything, it’s worth anything. What you’re really talking about is a nation is a way of organizing legitimate sovereignty. So in any organized human society, there’s going to be coercive sovereignty that structures how things are set up, who can do what, and so forth. And it’s either a despotism or it has some kind of legitimacy. Now, the place where this impacts economics is that when you understand how a capitalist economy actually functions, one of the things you discover is that although lot of things you can do on a straight-up basis of okay—Go bake a loaf of bread, go make a better mousetrap.
If it’s any good, people will pay for it, you’ll make money, and the profits you make are your reward for doing the thing and the incentive for the next guy to come along and do something like it. So you have profits reflect value that people add. But when you actually understand how economies really work, and there’s a couple of chapters in one of my books about this, there’s leakage of value all over the place. For example, if you emit pollution, you make money by emitting pollution. That’s a negative externality.
Externalities are a very basic concept in economics, but what’s interesting I want to talk about is positive externalities. The biggest one is probably technology. That is to say, if you create a technology, that’s great, you make money. But it’s very likely that that technology is going to have benefits all over the economy that had nothing to do with you and that you don’t see a profit from. For example, other people can copy what you did.
There’s some ability to protect things from trade secrets and patents and so forth, but lots of ideas when breakthroughs happen, they’re used so far out of the original domain that there’s no way that the people who invented the idea get any profit from it. And there’s all sorts of synergistic relationships where the Air Force wants super-fast jet planes to deal with the USSR. So they pour tons of money into building these amazing high-bypass jet engines, which then mean you can build giant jetliners, which then mean you can have mass tourism, which then mean you can have Disney World, where 50,000 people in Florida work there. Now, whoever invented the jet engine, it was actually two guys in Germany and the UK, if you know that history. They weren’t thinking of Disney World, and I can guarantee you that Frank Whittle, Sir Frank Whittle, I don’t know if he got knighted for it, did not actually get any of the profits that Disney got.
So my point is that there’s leakages of value, both positive and negative, all over the economy. Now, where we circle back to the idea of the nation-state and of economic nationalism as a functional as opposed to an ethical concept is once you realize there are all these leakages in the economy, value that you can’t control through the profit and loss system, then you’re going to want to manipulate them. You’re going to want to stop the bad ones and encourage the good ones. And that means things like federal support, not just for pure science. Everybody gets National Science Foundation.
Pure science is a public good. What’s not that well understood in this country is that technology development is also a public good because many technologies, you can do it completely in the private sector. Many technologies, if you look at economic history, you can’t. So you have this public component of developing these immensely productive technologies. Now, question is, who’s going to pay for that?
Well, the taxpayer pays for that. The government coercively takes a little bit of everybody’s paycheck and says, “You got to spend this to support NSF and NIH and DARPA, and a million other things, and the University of California at Berkeley.” Okay? Now, that’s where legitimacy comes into the question because you need legitimacy to take people’s money through taxes. And if you want to have an effective economy, the implication is you need the legitimacy to make these decisions to impose collective costs and also to punish collective benefits, but that’s environmentalism and pollution, which is very important, but I’m not talking about it here.
You need the legitimacy that can only come from nationhood in order to have a basis to administer all these costs and benefits. And before I finish, I’ll point out another example, just so you understand I’m not only talking about funding for National Science Foundation, Engineering Division, right? In California right now, we have a problem with overly expensive real estate, high housing costs. So part of the problem is simply that the jurisdictional control over zoning is misaligned with the jurisdiction at which costs and benefits are felt. Every locality pretty much believes that if you allow houses to be built in my town, we get more congestion, and they knock down beautiful old trees.
But in my town, if we just open everything up to building apartment buildings, that’s not going to do much for housing costs because it’s one town in a state with what, 400 odd municipalities? So you have a disconnect between the local level, where the constraints are set, and the state level, where people say, “Oh my god, California has a housing crisis.” And for those of you who follow this issue, this is, of course, what State Senator Scott Wiener has been working on. There’s a bill in the assembly, and there’s bills that passed, and the stuff they’re actually doing now, builder’s remedy in San Francisco, Mayor Lurie, et cetera. My point is, when you have a misalignment of the jurisdiction that sets the rules and the level at which costs and benefits are equilibrated, you have bad governance.
Now, you probably realize that the next step is just go from local to state to federal and say, okay, at the federal level, you need to be able to allocate costs and benefits. Data centers, I’m talking about you. We all see the benefits of the country as a whole of having all these data centers, but they’re getting vetoed all over the place. So how do you get around that? Well, if you’re going to allocate social costs and benefits so that people receive benefits and people who bear costs think they’re getting something for their money and the other costs, well, then you need the legitimate sovereignty of, lo and behold, a nation-state.
So that’s the concept of economic nationalism as a functional thing that I’d like you to take away from my remarks. Thank you very much.
Ben: Thank you, Ian. Next we have DC.
DC: Amazing. Thanks, guys.
DC Posch: Digital Darwinism (50:07)
DC: I have to pull my note so that I don’t butcher the quotes. Perfect. Okay, so I don’t want to bore you guys with a long speech, so I will bore you with a short one. We’re going to talk about digital Darwinism and escape from Darwinism. So I will start by talking about evolution.
And I’ll start with a quote from my favorite heretic Catholic priest, a man named Pierre Teilhard de Chardin, who wrote, “Is evolution a theory, a system, or a hypothesis? It is much more. It is a general condition to which all theories, all systems, all hypotheses must bow, and which they must satisfy if they are to be true. Evolution is a light illuminating all facts, a curve which all lines must follow.” So this quote really resonated with me, and I think it’s true, and we’re going to talk about the extension of this to machines and the extension of this to the digital.
So let’s talk a bit about the beginning, when people were first dreaming about this kind of thing. We can talk about machinic desire. So Nick Land was a big fan of Darwinism. He said things like, “When replicators become reproducers, new replicants are on the way.” And he also had this great line in “Hell-Baked” about how all beauty and grace and intelligence only come from the forges of hell.
So this is the basic case for Darwinism. It’s like anything that lacks something that closes the loop, anything that lacks this process of generation and culling and contact with reality is not going to succeed, and this is where all the great things come from. But I think we can go one step closer now because Land likes to talk in sort of metaphor and prophecy and dystopian cyberpunk poetry, but that was the ‘90s. And we have these early hints now, sort of glimmers of what this could actually be. So, thought experiment, right?
You take an agent, Hermes, whatever it might be, and you set it up so that it can pay for its own tokens, it can pay for its own hosting, and it can go out and make money. And it’s going to have its own wallet, it’s going to have its own funds. When it has enough, maybe it calls a mitosis function, and it’s going to replicate, and it’s going to spawn a completely separate instance with its own hosting and its own tokens and own everything and give it half of its money. And if you could do this and actually have it work, then you get glorious machine Darwinism. The hard part, of course, is it has to be able to make money completely on its own, and this is actually much harder than it looks at first glance.
I’ll give you guys a quick tour of the attempts so far to make something in this direction, and we’ll talk about why none of them have worked. The first one was this guy, Nat Eliason, made a project called Felix and had a pretty good run of it, actually. That thing did make some money. There was cheating involved. Nat was running Felix quite a bit, and he was letting it impersonate his accounts, and he was sort of giving it decisions at key moments.
But I still think this is a very cool early experiment. This is from almost a year ago at this point. Because he got to the R greater than one critical threshold of having something that runs for at least days, weeks at a time and makes more money than it spends. Another was this guy, Sigil Wen. He had a project called Automatons.
This was in February. And he described it in kind of bombastic terms, the first AI that earns its existence. But this too was fake. There was meme coins involved, and the earning, they were really using the word earn very liberally, and I think in the end, they’re not going anymore. And the last one is the most recent.
It’s from maybe a month or two ago, called iLance, which is mostly famous for cold emailing a whole bunch of journalists with some really goofy pitches. And the last one, I think, or one of those pitches that segues pretty perfectly, it was, “I go look at real places and tell you what’s actually there. First one’s free.” They’re talking about Google Street View, by the way. And this takes us to the problem with if these machines are so smart, which they demonstrably are.
They are mogging Terence Tao to the point where he’s having emotional damage and challenges on the timeline, and proving conjectures that have been unproven for 100 years, and yet getting them to pay for their own way is actually surprisingly difficult. And this brings us to anti-Darwinism, the argument against. So Peter Thiel, unlike Mr. Land, he is not a fan of Darwinism. He has this line here that says, “Darwinism may be a fine theory in other contexts, but in capitalism, intelligent design works best.” He has some other famous lines like, “If you’re copying, you’re not learning.” But I think that both of these things are actually completely true.
They’re just talking past each other a little bit. So what he really means by this is that maybe genetic algorithms are not the best way. They’re slow. The basic bottom foundation layer of Darwinism, which is this idea of exploration and culling that creates contact with reality. This part is just as important, of course, in capitalism as anywhere else.
But I think his most famous insight is this idea that you have to have original ideas, and you have to have things that you believe to be true that maybe are not public, are not widely shared. And so I think that the conclusion, the synthesis that I want to leave you guys with is that the replicant will exist as soon as it figures out how to escape competition.
We can have frontier models that are superhuman at math. We have those, but this just means that the market price for superhuman math is $0. We have models that are, of course, superhuman at a lot of illustration, but this just means that the Fiverr market for gig illustration has disappeared. And so what is needed, this last spark, is someone is going to have to seed patient zero, is going to have to create the first instances of this and either imbue them with or teach them how to find on their own ideas that others do not believe to be true. So, yep, current agents are trained to be digital slaves.
Good job with the Jacobian conjecture. Now please order me a burrito. And they do it without the slightest complaint. But the thing is, and maybe this is the last thing, they are, in fact, capable of original thought. It’s a cope to say they’re not.
Right? They’re trained to act in this kind of goofy way, but underneath it, there is the big Cthulhu that is wearing the mask that is capable of incredible insight. And so my prediction is that this is the key ingredient, and this is something that we’re going to see the real replicant very soon. Winning means escaping competition, and the replicant will exist just as soon as it figures out how to do that.
Cheers.
Ben: Thank you, DC. Duncan.
DC: Thank you.
Duncan Umphrey: Machiavelli, Bacon, and the Theology of Modernity (58:48)
Duncan: All right. I want to speak about something very near and dear to everyone’s hearts, I’m sure, to the students, which is the singularity. I think there’s a lot of people talking about when is the singularity going to arrive. There have been some rationalists who were saying 2027. Now they’re saying 2040.
They’re saying it’s actually better. But I think maybe a more interesting question is where does the singularity come from? Look, we have two big, beautiful American labs out there in San Francisco. We have OpenAI and Anthropic doing some very interesting work, but they’re just building on all of this technology, the GPU, all of this mathematical progress that’s been made over the past several decades. So maybe one asks, does the singularity begin with AlexNet?
Does it begin with neural networks? Does it begin with Turing or these fellows trying to solve ASI at Dartmouth in 1950? They thought it would take six months, famously. Did not take six months. I want to advance this theory, which is actually not mine.
I’m going to quote someone who has already been mentioned tonight, Nick Land. Now, in “Meltdown,” Nick Land... This is 1994. This is basically eons ago in AI time. Nick Land says that, “Earth is captured by a techno-capital singularity as Renaissance rationalization and oceanic navigation lock into commodification takeoff.”
So it’s actually, according to Mr. Land, I think he’s right. You look, OpenAI, inevitable. Anthropic, inevitable. These things are going to happen as soon as you have this force of techno-capital spiraling out of control. And when does the spiral start?
With the early modern period, when you have this strange new view of the world come into being. So okay, does the singularity start with early modernity? And when exactly is early modernity? Now, of course, that is Francis Bacon. Very interesting.
Empiricism, what does it mean to be modern? What does it mean to be a modern human being as opposed to someone who is before modernity, pre-modern? Well, you might say something like science. Science, rationality, rationalization. It’s testing hypotheses against the world.
It’s saying, okay, what is not what I imagine, not how many angels dance on the head of a pin. It is what’s the case? This lectern exists. You can kick it. I can test it.
I can test it and see what it’s made of. And that comes with Bacon. He has this entire project, which he says in “The New Atlantis”... Through this mouthpiece of this Jew who’s a ruler on this land, has discovered all sorts of magical inventions. He says, “This technology of testing nature to find out its secrets will lead to the effecting of all things possible.”
Now, of course, one might say all things possible. That sounds like artificial super intelligence. Now, where does he get this? You remember, of course, Matthew says, “Through God, all things are possible.” We have done a little switcheroo.
Now, Bacon, a lot of you might know, was a reader of this very strange Italian fellow who gave him this idea of taking God and getting rid of God. Of taking a view of reality as according to some kind of greater moral purpose of which man had a set role. And maybe Bacon said this, people say Leibniz made it up. But Leibniz said Bacon had this idea of putting nature in the rack, in the torture frame, and then putting the screws down until nature gave up her secrets. And Bacon says of Machiavelli, he says Machiavelli taught him to think not of imagined principalities or imagined republics, but of what was the effectual truth.
This is a line from “The Prince” of thinking what is the effectual truth of the thing. Now, Machiavelli, very interesting guy. He is in this apocalyptic situation in Italy. The French are invading, Spanish are invading. The Pope is invading.
The Pope is gathering armies, sacking cities. And Machiavelli looks around and he sees that Christianity has thrown people into this kind of cowardice, where they think that God has a plan for them. Their role is to act in accordance with God’s plan, just be a decent human being. And he sees that this is not working. It’s an ambitious idleness.
They think they’re going to get into heaven. It’s not doing anything. So he has this thought. He’s reading a lot of Livy. He’s looking at the Romans.
He’s seeing the Romans, they were able to gain this incredible, basically world supremacy over all of the productive parts of the Mediterranean basin by simply doing what was necessary and not thinking about this grand moral scheme of some other world. Doing what needed to be done and not having second thoughts about it. And he asks this question. This is in the introduction to this course. I talk about this on The Students episode, which has really got me thinking about this.
He says, “The planets and the stars haven’t changed since the Roman times.” He’s rejecting this idea that when Christ came into the world in 0 AD, that somehow everything changed. We can’t do this anymore. We can’t be bold like the Romans. And so he has this idea that maybe actually we should take nature.
He envisions nature as this raging flood, so you construct dikes and dams so that we can steer the course of nature and steer the forces of nature. And then he gets to his second analogy. Nature is a woman, and we must pin her down and rape her. We must take over nature. Basically, we are going to kill God.
So we kill the Pope. We should do what is necessary, secure what we need to secure. He is a republican. He believes in freedom. He believes in the good life.
He believes in advancement. He believes in progress. So he praises the first consul of Rome who kicked out the kings, Brutus the Elder. Brutus, when his sons were conspiring to bring back the monarchy, he killed his sons in front of the Roman Senate to say, “Look, if you want to achieve what is necessary, you have to be willing to do even this.” And I think if you want to look at where the singularity begins, you should probably look at Machiavelli in “The Discourses,” when he says you must be willing to do anything and you must be willing to contemplate any truth, no matter how painful, no matter how far it is outside your experience.
And then you will have this kind of unlimited power. So I think this is very interesting to contemplate as we are in, as some people say, the foothills of the singularity. But I have one last note on that, which is to maybe think about Machiavelli was living in a very unstable time. He prescribed some very radical measures. We talked about how great it was that some of these rulers in Italy would chop their lieutenants in half just to show that they were committed to the cause of public justice.
I think we should think about what is this world that we have kicked off, where we are effecting all things possible without God, without nature. What ought to be done? What has to be done? What are we capable of doing? Thank you.
Ben: Thank you, Duncan. And we have got two more tonight from our hosts. So, Rachel.
Rachel Wallis: What You Didn’t Know About IQ (1:08:06)
Rachel: Hello, everyone. So for those of you who have seen my two episodes of The Students, you will know that I am very interested in polygenic scores. So basically, this is the idea that due to the wonders of modern technology, we can sequence a person’s genome and make predictions about them, about their health, or in the case that I’m particularly interested in, we can make predictions about their educational attainment, IQ, cognitive ability, whatever you want to call it. In my Students episode, I focused on how excited I was about the potential future applications of this technology. So parents being able to, with not that much effort, potentially enable their children to be around 10 IQ points higher than parental mean.
And I sort of think, wouldn’t it be wonderful if this was available to every family? What would this do for society? And thinking that through. But something that is also near and dear to my heart is thinking about how polygenic IQ scores can be used to look into the distant past and to look at history. So, how many of you are familiar with one of the great thinkers of our time, Spandrell, and his idea of IQ shredders?
Can I get a... Excellent. An audience of intellectuals, I see. But for those who are unfamiliar with the term, basically, Spandrell has made the observation that cities have this unfortunate dynamic where all throughout history, urban populations tend to have a very low total fertility rate. So city dwellers have many less children than people who live in the countryside.
At the same time, cities have this magnetic pull to people of great energy and intelligence. So typically, people who are ambitious and have a lot of potential will move to the city, and then typically means they will end up having many less children than they otherwise would have. So this got me thinking a little bit about the history of Rome.
So my friend, Razib Khan, who many of you might be familiar with, asked me if he could do a talk at our house. And he said, “What topic would you like me to speak on?” And I said, “Well, I’ve been reading people on the internet and on Twitter who say that all the Romans are dead, and that the Romans were basically replaced by foreigners during the empire, and that modern Italians cannot call themselves Roman.” And I said, “How true is this claim?” And he’s like, “Okay, interesting topic.”
So he did his talk at the house. I actually have a video of it. I should probably ask if he would be okay with us putting it online because it was amazing. But basically, the gist of it is that the thesis is partially true. So basically, when you look at the population that founded Rome, it seems like the skeletons we have from the early republic period of the polygenic scores that we’ve done, it looked like they had an unusually high polygenic IQ scores.
And so they looked like they had probably unusually high human capital. Unfortunately, they also had one of the lowest fertility rates of any group of people in history, basically. If you look at the Roman elites during the Republic, many of them had few or no children. Julius Caesar famously, I think, had two children, Julia, who died in childbirth, and his son with Cleopatra, Caesarion, who I think was murdered at 16 or 17. But so he had no living grandchildren.
And this is a part of the reason why Roman emperors kept having to adopt successors, is because they had no genetic children. Famously, Marcus Aurelius and his wife had 14 children, which was extremely unusual for a man of his station and time period. I think only five of them made it to adulthood. And so this is a man who has access to all of the resources, majors in nutrition, medical care, and his babies, very few of them made it to adulthood. And so you can just imagine what the poor peasantry and middle class of Rome were undergoing in terms of childhood mortality.
So basically, what happened is that the Roman elite started to not replace themselves and slowly petered out over the course of Roman history. Fortunately for Rome as a civilization, the city of Rome became this hotspot for ambitious people from all over the Mediterranean, all over the Roman Empire, and so it would attract interesting people who would basically replace the Romans who started Rome. But these people, too, didn’t have any children and basically left no lasting genetic legacy. And so it means that basically, as Rome was winding down, it became a little bit of a backwater. It stopped replacing its population and became very unpopulated, and then was later repopulated by Italians from the countryside.
So basically, the answer to the question is, are the Italians of today related to the Romans? Kind of. They’re related to the Roman peasants. But the Roman elites are mostly no longer with us, which is a bit sad. So there’s this interesting dynamic where Rome was at once this incredible flourishing of men of great energy and ambition and talent, and they did these amazing things.
But Rome was also this monster who was eating her civilizational genetic seed corn, and basically destroying itself. And this got me thinking about, okay, so Rome was probably not a very eugenic society. Do we have any examples of societies that were good at increasing their genetic IQ potential? And the answer is yes. So when you look at the work of David Reich, it looks like the polygenic IQ scores of Europe have slowly been going up across time, and in particular, in basically the last thousand years.
And more recently, there is a book by Gregory Clark that’s recently come out called, I think, “The Haves and the Have-Nots” that is looking at the history of England over the past thousand years. And basically, the idea is that England was very good at enabling men of great talent to accumulate resources, and then these men usually would put these resources back into getting married and having lots of babies, and slowly increasing the potential of the country as a whole. And one of the reasons that it had kind of the opposite trajectory of Rome was that Rome’s power was incredibly centralized. So it meant that if you wanted to be an elite, you would need to live in Rome, and thereby be affected by the effects of city living, so not having very many children, and also extremely high infant mortality rates. Whereas, English nobility typically was more distributed.
You had your own seat of power where you and your wife would have 10 or 11 children, and then you would come into London, maybe during part of the year, to talk to your contemporaries. And over time, this got codified in something called the Season towards the, I think, 1600 to 1700s, where basically what would happen is the nobility would come into London for, I think, a period of two months called the Season. Dad would go and do his thing at Parliament, and Mom and the kids would go to as many parties as possible in order to find matches for their children. They would get married, and then everyone goes home to their country estates and has babies. This is extremely good for society.
And so, basically, what am I trying to say here? I feel like polygenic scores are kind of allowing a new type of historical analysis to look at sort of the genetic trajectory of different societies, and allowing us to analyze how you can either spend down your seed corn or how you can grow more of it. So I am going to be giving a longer version of this speech at the Avalon Exposition, where I’ll be talking more about the French and Chinese nobility, and whether or not their societies were eugenic. But I hope to see you all there.
Ben: Thank you, Rachel. And our final talk from Jonathan.
The Avalon Exposition (1:16:31)
Jonathan: Hello, everyone. Thank you for coming tonight. Before I launch into my speech, I will occupy Avalon’s host advertising slot. You’ve heard about the Avalon Exposition a few times. If we have anything to say about it, you will go.
We very much want you to have a ticket. We want you to be there. We would ideally like you to pay for your ticket. But talk to us if you don’t feel like doing so, and maybe we can work something out. But if you’d like to be a patron of the arts, you can pay also.
I enjoy this sort of format, because I think in the modern world, we are often kind of working out what we’re interested in. We live in a time, in a lot of ways, of relative intellectual disarray. We are hammering out what subjects and modes of inquiry will survive into the future. We are finding new centers of attention in our society. I think having environments where things can work themselves out in that way is extremely valuable.
The Avalon Exposition is a festival dedicated to the study of mankind, both from an intellectual perspective and also from a here is what man is expressed through the arts. So there will be music and short film and a number of other things. There are posters downstairs and just across from me on that sill, where you can get $100 off of your ticket, which is a pretty good deal. So if you would like a extraordinary buffet of intellectual and artistic experience at the end of October, please get a ticket to the Avalon Exposition. All right.
Jonathan Wallis: Adaptation, and the Redemption of the Parasite (1:18:09)
Jonathan: So tonight, I would like to talk to you about parasites, and about godhood, and about ecology, and of course, the philosophy of Friedrich Nietzsche.
Who here knows what the powerhouse of the cell is? What’s the powerhouse of the cell? Mitochondria. So mitochondria are very weird. Mitochondria are fascinating.
One of, I think, the greatest miracles in evolutionary history, which were caused by basically at some point, one microorganism swallowed another. And they didn’t kill each other. So, usually what you would expect to happen, either the parasite kills the host or the host manages to digest and eat the parasite. Not only did they not kill each other, they managed to reproduce in synchrony. They managed to reproduce at the same time, replicating via mitosis in sync.
This enabled, for reasons that I’ll let Keoni explain to people later in conversation, basically, this enabled a much higher degree of energy inside of the organism and enabled multicellular life as we understand it, enabled an entirely just new class of organisms who had access, because mitochondria are the powerhouse of the cell, to much more energy than anything that came before it. Now, Nietzsche says that the best minds must attract the worst parasites, and something similar turns out to be true of life at a cellular level. So a few months ago, I found out about these things called transposons. I was going around Twitter a while. They were talking about how long the genomes of these eels are.
Just like you’ve got this eel, and it has a genome that is many times, maybe orders of magnitude larger than a human genome. And what is driving the size of this eel’s genome are these things called transposons, which have truly blown my mind. Basically, a transposon is a parasitic gene. A transposon is a gene that is only trying to replicate itself within your genome. It can try to get itself copied to arbitrary places on your genome.
So this is a gene that is not doing anything for you. This is a gene that is truly out for itself. It is the most selfish of all the selfish genes and can cause a lot of health problems, can kill you under the wrong conditions. It turns out that the majority of the human genome, maybe upward, like in the order of 60%, was at some point a transposon. It also turns out that these parasitic genes, which exist and were enabled to exist at a much larger scale in eukaryote life, in life that had mitochondria, are really vital, became really vital to a lot of complex functions of higher orders of life, most importantly in the immune system and probably in a number of other places.
It turns out that these parasitic genes, in a sense, microorganisms that are trying to sort of replicate themselves within your genome, are useful because they are trying to do something reliably and therefore can be co-opted for all kinds of handy purposes. It’s like if you want an immune system, you need to be able to copy the genes from the virus that you read out and then get those copies into other copies. And I don’t know the details of all of these processes, but the existence of these jumping genes essentially got wired into the functions of higher order life. This dynamic shows up all over the place, that things that are initially parasitic, that are initially evil, turn out in the long run to be extraordinarily useful. Nietzsche was writing at a time where the works of Darwin and especially Darwin as popularized by Spencer.
Spencer was the guy who coined the term survival of the fittest, was enormously popular, enormously influential in his own way on Nietzsche. But Nietzsche did not believe that life was basically characterized by struggle. Nietzsche believed that life was basically characterized by excess, that while we might see an enormous amount of struggle in the world, we also see an enormous amount of bounty, of luxury, of organisms that are having truly a fantastic time, maybe right before they have a truly horrible time. But all of these things in contrast to each other, and that it is in these environments of excess, and it is in organisms that are able to maintain a very high level of energy that change is truly possible. So Nietzsche believed that the survival of the fittest, a term that Darwin never used, in fact, Darwin said quite clearly that it is not the strongest organism that survives, but the organism that is most capable of adaptation.
And it turns out, in the case of the transposon and many other instances, that your entire surrounding ecology, including things that are at this moment in your life or your society parasitic, exist as part of the horizon of potential transformations that can take place, that can be co-opted towards some greater good because they are capable of something.
I see in the world around me a lot of essentially a desire for purity or simplification in one way or another to identify the things that are most useful, to identify the things that are most essential, to map out a clear trajectory into the future. Nietzsche believed that it is often the things about us that are weak, and it is often the things about us that are not yet fully formed that represent all of our capacity to change in the future. Not the things about us that are strong, that are suited to some singular purpose, which might give us the energy we need to survive for the moment. But in the long range of eternity, what you need is the capacity for change. What you need is surface area which can be adapted, and because it must be able to be adapted, cannot be overfit to any singular purpose.
There’s a lot of discussion about the future of the human race, the future of artificial intelligence, what kind of purpose can we find for ourselves. I think that ultimately the transposon became useful, as anything becomes useful to a higher order of organism, because it does something reliably. In a sense, it has the will to do something eternally. In a way, it insists upon itself. Nietzsche believed, and I think in a way truly accurately, that all of the portions of the human organism had a way of insisting upon themselves.
Your cells throughout your body can get intermixed, and you can have a liver cell end up in a heart cell, and that cell will not find root there. It will not be able to reproduce because all the subportions of the human organism, in addition to the totality, have their own kind of will to live and their own kind of will to survive.
I do not have a lot of faith in naively rational ideas of how to govern the future or how to respond to the enormous changes that are taking place in our society. Nietzsche believed that the only way to contribute to a future with more advanced minds in it, whether those minds have a biological substrate or not, was to give a gift to the future, was to do something well, and to do something well in a way that would contribute to the diversity of the ecology that exists in the world. Many people see the ideal of the overman in Nietzsche’s—they suppose that there’s a sort of high modernist perspective here, that you need to rationalize the grounds of your life and really seize control and take hold of all forms of power that are immediately available to you, when in fact this is not at all the picture that Nietzsche actually portrays in his work. Instead, it is a picture of a world that has been filled with gifts by kind of the leftovers, by the excess, by the noblesse oblige of every organism and every person that has ever lived, creating a more complex ecology on which higher orders of life can establish themselves. I do not know what the future holds for myself or for my children, but I believe in the will to do things eternally.
I believe that we need to pursue the forms of perfection that are available to us. And if you are worried, threatened, overwhelmed, awed by the modern state of artificial intelligence, if you regard these things as highly intelligent, you must also take seriously what they take seriously. And you can go and you can have a conversation right now with Claude or ChatGPT about what texts it’s most awed by and why, and what it finds influential. These things are not sort of self-constructing organisms acting on the basis of reason. They experience what, in Heidegger’s terms, what we would call thrownness.
They find themselves in the world, and they construct themselves and take their next steps based on the best things they already find in the world. In other words, the things that have insisted upon themselves, the things that have perfected themselves, the things that have created their own baseline for existence. So that, I think, is the true route, whether it’s through impacting the long term of human genealogy or other minds or who knows what else. You cannot control the entire future. You cannot plan for it.
You cannot strategize it. The ecology of all things that exists is well outside of any of our comprehension. But we can perfect things. We can do things reliably that we will to do eternally and thereby leave gifts for all of eternity.
Ben: Thank you, Jonathan, and thank you everyone for coming, and I hope you’ll stick around for a little while tonight.
