The Exchanges

Q And now you made a lot of money, and you're stopping, you're thinking, how do I have to go back and just to fight for society again in different ways?

A Yeah, and I had my second of two kids, and that was huge, because I was trying to think, what do I model for these little humans? You know, what's the world they're entering? And so I started reading, like, 15 minutes a day after I put them down, and that grew to, like, three or four hours a day. And I had a really good mentor, Michael Strong, who you know, he wrote me a 17 page annotated syllabus, started with the ancients, moved all the way forward, and it just kind of took over my life. And the first insight I had is there are more people like me who are tech entrepreneurs, who studied STEM or business or something like that.

AI assessment note: “I had my second of two kids, and that was huge, because I was trying”

Q And so you founded the Cosmos Institute. What's the Cosmos Institute?

A So we're trying to, we're, we're an academy to try to cultivate these philosopher builders. So we think that that's a key archetype that we need, and I can kind of, you know, the historical thing I think you'll appreciate, because, you know, you know and think about a lot of these characters. So my contention is that whenever you have a technological revolution, you have two choices. You either build for freedom, or you watch as others build for control, and I'll give you examples of this. So Benjamin Franklin in the US, he builds Uh, a network of libraries and independent publishers that takes the printing press, innovation, and realizes the democratic ideal of enlightenment, sorry, of spreading knowledge. At the same time, you have Joseph Goebbels, who takes the printing press and turns it into a system of mass manipulation. Forward to the Industrial Revolution, you have people like Adam Smith, Theorizing about markets, creating broad prosperity. You have the Soviets taking industrial technology, twisting it into a system of economic domination. And then lastly, in the internet revolution, you have people like Tim Berners-Lee, who embed decentralization into the technical fabric of the world wide web, give us an expressive open internet. And you have China creating the great firewall and the social credit system that creates this web of conformity and control.

AI assessment note: “we're an academy to try to cultivate these philosopher builders.”

Q global kind of frameworks to take away our autonomy. And obviously to a lot of us who believe society should be free and work bottom up, this is like one of the worst things ever we'd fight to stop. Like, like what, what's the history in the enlightenment or, or, or more recently with this world government idea? Cause it comes up a lot in a lot of this work.

A Yeah. So one example here is when, when Alexis de Tocqueville comes to America in the 18 thirties or, you know, around 1830, he goes to New England townships and he sees something that's totally different from what he had seen in Paris. Paris is very hub and spoke. Like the, the, the government in Paris controls, you know, controls France at that time. And in America, in New England townships, you see something totally different, which is bottom up people associating spontaneously solving problems. They need a bridge, they go build a bridge. And it's this incredible release of energy. And Tocqueville thinks that this is very important. The reason he thinks it's important is not only because, like, of the vibrancy, but because he recognizes that in democracies in particular, the individual is weak. The individual versus the state is not a fair match. If you're a serious aristocrat, maybe you have a, you have a chance, but individuals in democracy are weak. They're also weak relative to other forces like The will of the majority. And so Tocqueville says it is essential that you have a decentralization and a vibrancy that comes from the bottom so that they can fend off these powerful forces. If they do not, then they'll end up with encroachment, they'll end up with what he calls soft despotism. We could talk about that. More, more recently, I think, you know, one of the most inter…

AI assessment note: “one of the most interesting debates on this issue is Leo Strauss versus Kojev”

Q I want to jump to just high level for people, the various approaches right now that dominate the AI discourse, right? So there's, I think you define these, there's the doomsayers, there's the accelerationists, I think there's the regulators, and there's techno-authoritarians, and there's a four list. So what do you think of each? What are the doomsayers, first of all?

A Okay, so that's the most, um, that's the intellectual incumbent, um, I would say. That's the most common, uh, dominant philosophy in the practitioner community, and it's three philosophies in a trench coat, but they're all very entangled. One is effective altruism, One is rationalism, one is long-termism. So three things, you might hear these terms, but you can kind of put them together and say those are, like, the doomsayers. And the gist of it is that AI could pose a risk to humanity, and we should pause. It's kind of like Oppenheimer when he had this, you know, hubristic awe of his own creation. Or you could say, like, a tragic kind of Prometheanism where, like, we got fire from the gods and now we regret it or something like this. And so that, that school tends to then follow, like, One of the main equations, people get mad at me for saying this, but this is actually the way that a lot of folks think about it, is even a tiny probability of doom multiplied by negative infinity utils equals negative infinity utils.

AI assessment note: “three philosophies in a trench coat... effective altruism, One is rationalism, one is long-termism”

Q Tell me more about that. What's, what's immoral about the doomsayers?

A Well, so, I mean, the, the key premise is that moral questions can be reduced to a kind of, to a thing that's called a util, or to a common currency. This is, this is something I would reject. In other words, I think that endemic in morality is the idea that you have a bunch of good things that are incommensurable. They can't be calculated, like courage, you know, courage in battle, familial love, religious observance. How the hell do you compute those things? You, you can't. It's actually sharply heterogeneous goods. And I think because you can't do that, the, the sidestepping move that a lot of these folks do is they say, well, the thing we really care about is human life. We're just going to focus on human life. It turns out there are a lot of other human goods than mere life that we should be focused on.

AI assessment note: “the key premise is that moral questions can be reduced to a kind of... util”

Q Basically there's lots of context and pairing of context in order to make it work. So we have these pillars, the AI and human flourishing, like what are, what are you doing to apply these principles? What's the goal? How are you working?

A Okay. So the big thing here then is that, um, I talked about how Franklin and Smith and Tim Berners-Lee made a huge difference, right? In, in having technology bend towards freedom and away from control. And so the conclusion from that is we've got to cultivate people like that. And I call them philosopher builders, people who are on the one hand, willing and able to contemplate deeply about the ends of the technology. And the other hand, they can actually build stuff. And I think about like, do our institutions, uh, prepare those people today? And I say, no, And I say, like, you start at universities, right? And I'm gonna make a general statement. I'm a big supporter of University of Austin, which I think is an exception here, but they train narrow technicians or conforming ideologues. Tech companies train people that are very good at optimizing the means, not very good at deliberating on the ends. And then think tanks, theorists that don't build. And so I look at the, the major institutions that should be producing these kinds of people when we need them most, and I say they're not there. So then I say, okay, how can we How can we do that? How can we cultivate these kinds of people? And I think partly it's education. Like, I think there needs to be a course, by whatever, you know, formulation we say that, that fills the gap, where if you study computer science, or you study b…

AI assessment note: “we've got to cultivate people like that. And I call them philosopher builders”