Fireside chat | Tyler Cowen | EAG DC 2022

Tyler Cowen of George Mason University. Tyler, thanks so much for joining us. If you have questions for Tyler, you can submit them via the Swapcard app, and I will look through them and try and read out some of them in the course of our conversation. So I was up here yesterday with Michael Cramer, and his pitch to effective altruism is that you focus more on innovation, or specifically on things where, like, their social value is underrated by the market, like advanced commitments for vaccines against diseases that disproportionately affect poor people, pesticides that are stuff in Africa. And I was taking very careful notes for this talk, because I was like, I bet that this is like interestingly close to your perspective, but also maybe a bit different. Because one of the slides he put up here was comparing child mortality in 1910 to now by GDP. And the point he was making was that at a given GDP level, child mortality is now like 40 times lower than something than it was in 1910. And so that's not like general growth and like wealth and prosperity. That's like specifically that we like fixed a bunch of stuff that was killing kids. And I was interested in like what daylight you would identify between like your worldview about how to make things better and that sort of view.

I'm not sure I know exactly what Michael's view is. I certainly think innovation is one of the most important things, but it's striking to me how many countries in the world are not close close to the innovation frontier. And there's simple things they could do, like, say, supply clean water that would also help child mortality. And they don't do because of bad institutions. So the relative weight he and I would put on institutions versus innovation maybe is different.

What do you think of as like the most important things that something like EA should be doing or that countries that are trying to grow should be doing?

Well, what EA should be doing and what the countries should be doing are very different questions. I think of EA as a bit too much taking an optimizing perspective on what is the most valuable thing to do. You know, is it anti-malarials? Is it bed nets? Is it deworming? Is it whatever? I am much more likely to start with the question of who is the individual and what is that person good at and what are their passions? And to build out from that, I think you get higher productivity than to try to solve in the abstract for what is the best thing to do. So I see in a number of people like I've only been here an hour and a half and already 5 random people have exhibited to me what I call moral nervousness. Like, am I doing the very best thing? And if not, you know, do I need to worry? I joke to the people back there, can I leave my bag in the back? And they say, of course. I say, well, shouldn't you steal it and send it to Africa? Right. If you're morally nervous, but you shouldn't. So In that regard, I have a different perspective than EA. Mine is very person-centered and from the individual outwards.

I'm curious, like, how— like comparing something that like Emergent Ventures has done or something that you've spearheaded that's like done a ton of good in the world to like the stuff that EA does. Is that like coming from your different perspective about person-centered approaches? Is it coming from like a different philosophical difference between you and EA? Or like when you launched Fast Grants, what were you thinking about? Like how this is going to do good in the world?

With Fast Grants, that was a program set up early in the pandemic to accelerate the pace of biomedical research because the NIH and other governmental bodies were failing so miserably during an emergency. They were still taking like 5 to 9 months or more to referee proposals and get the money out. So there the goal was just to speed up the end of the pandemic. In general, I'm not sure I'm the great expert on like what EA does. I suspect in actual practice, when you actually have to run a program, the difference between what EA does and what non-EA does is somewhat less than one might think from rhetoric that you want things to succeed and you end up targeting talented people in any case. So for all the talk on EA forums about like how many chickens equal a cow or is deworming better than the anti-malarial, whatever, when it comes to actual practice, I'm not sure that's shaping the allocation of resources that much. It's like, who's on the table in front of you? Like, here's Kelsey Piper. What are you good at? What do you know the most about? And that's what you end up doing. So in that sense, the differences probably are quite overblown.

I actually am kind of with you there. I also have a bunch of meetings at EA Global with people who are like, what is the thing for me to do? I was like, I just met you, you know? Yeah. When I was in college and a baby EA, I was like, well, at the time, Earning to Give was sort of the focus. Making lots of money to donate. I was like, I guess I should be a programmer. Programmers make a lot of money. I was a terrible programmer. I'm like, I'm good at writing stuff. And I do think that like EA is making a mistake to the degree it's like pointing people at stuff that they're bad at instead of stuff that like they can excel at.

But there's a chance EA is second-order efficient. Like this is a big meeting with a lot of really smart people so that you make the meeting about something else, some kind of tortured exercise about, you know, moral optimization, maybe you attract more talented people than if you just make it about talent. I don't know. But EA has been super successful. So in that sense, I'm not so inclined to second-guess you all, right? So like people say, oh, EA is a religion. It's not exactly true, but the elements of religiosity, you know, probably are part of the secret formula. So as I've said before, right now EA is the place with the most impressive, intellectually speaking, young talent of anywhere, any movement, anything. And think long and hard about how that happened. Right. But be Straussian about it. You know, don't think that what made that happen is necessarily literally true.

I want to talk a little bit about the future, like in the next 10 years and stuff. A lot of EAs are very worried about AI. Are you worried about AI?

Well, I'm sort of worried about everything. I'm not worried about AGI. I'm very worried about AI in the sense that an evil foreign power might take powerful AI and use it to start a war the way, say, Hitler used more advanced tanks in the Luftwaffe. That's by far for me a bigger worry than Skynet goes live. And if Skynet goes live is possible, the other is possible too, and the other will come first. But I'm much more of a historicist, I think, than what I take many people to be. I look at history, I see how civilizations fallen. Well, there's foreign conquest, there's pandemics, there's environmental problems. And those are my priorities because I've seen those happen many, many times. I've never seen a thing like Skynet goes live. So for me, it's a hypothetical. You can't dismiss it. But the powers of pure reason I view as fairly limited. So it's like trying to argue, well, what if aliens come from outer space and conquer us and enslave us, kill us? You can't say it's impossible, right? You could reason your way into being terrified. Well, we start with the Drake equation, and then I read Robin Hanson on the grabby aliens and so on, and wake up in the morning quite scared. But at the end of the day, the historicist perspective that pandemics, war, and environment are what have mattered and sort of stick with what you know, I still think is wiser.

So it still seems like if you think that, you know, a foreign power could develop a like quite powerful AI system and start a horrible war of conquest with it. Like, that's, that's pretty worrying. Like, that would be very bad if it happened. What do you think, like, should be done about it? Like, how do you make sure that AI development goes well and not badly if it's really important and potentially really valuable but has those huge risks?

I don't think you can make sure. But my first question would be like, is the evil power China or is it us? Now we could debate that. But if you decide on average it's more likely to be China, you want our AI to be really good. So that makes me more of an accelerationist because you're not optimizing for the whole world. You're like at best optimizing for one country, one institution. And if you think on average you're better than the bad guys, which again you can debate, but assuming you end up in that position, then I say full steam ahead. But that's not always the popular view in EA circles.

So it does seem like a pretty unpopular view in EA circles. I guess the first thing I would say to that is that like If AI has the potential to go well and be really good, and doing AI wrong has the potential to be really bad, even if like specific scenarios seem implausible to you, then full steam ahead seems much likelier to get us the bad outcomes than the good ones compared to like a more restrained pace. Like the fact that we developed nuclear weapons like in the middle of a war to deploy in a war seems like it maybe had some long-term destabilizing effects on how nuclear weapons like got used. Compared to if we'd like, you know, peacefully develop nuclear power, you know, over a longer period of time, not as part of a secret government project for a war.

But we had rivals then. So there were the Nazis, the Soviets. In my view, we needed to make sure we were first. We didn't necessarily know how far along they were. Very glad we had the Manhattan Project. And I mean, I would take a kind of sociological perspective on this also. If you look at the people who are most worried about AI risk, they're the people who are working the hardest on AI. And you could say, well, it's because they're the biggest experts, but they're the ones who are really going full steam ahead. And I take that observation, that correlation very seriously. I think maybe that's more important than the rhetoric. So EA as a whole is doing more to accelerate AI than, say, I am. I'm not really doing anything at all to accelerate it that I know about. And EAers are doing so much. So let's like look that square in the eye and just accept it. Like, that's what you all in the aggregate, weighted by AI expertise, actually do and think. I would say I don't mind that, to be clear, not a criticism. I would just say it's a reality. Get used to it.

That, that is one of the most spiciest takes I've heard up here. I think a lot of people will be kind of sad about it. But, you know, so—

but I've met with some of these people, visited some of the institutions. Not sure I should say which ones they are, but they're very well known. And you hear across the same lunch table talk the worries and the gung-ho-ness. And like, what's, you know, as an economist, you say, well, look at people, look at what people actually do. And that's the gung-ho-ness.

Yeah. So I think there is a lot of people who are like, I'm building AI and of course I'm going to make it safe because I understand the safety risk. I'm going to fix that. And then I think there's a bunch of other people who are like, That's a terrible plan. That's like going to be catastrophic. And I sometimes they're the same people. I think often the second group would say that the first group is like paying lip service to safety while they like try and build something that's not going to be safe. I don't, I don't know that like there is the people who are most worried about AI risk, I would say, are like mostly not the people that like DeepMind or OpenAI who are like putting out, you know, GPT-4, 5, 6, 7, 8, 9, 10.

I think sooner or later, usually sooner, there's an unsafe version of every innovation. So I don't know what it means to shut down AI work. To me, it just means China has it all, or maybe other foreign powers over time. And you could put in all the different safeguards that people talk about. But at the end of the day, the AI itself will get copied. The economies of scale will fall. The price of doing it will fall. And maybe you postpone Skynet goes live by 17 years. But if you really think that's such a risk, you won't postpone it very long and it will happen. So I treat that as like the aliens coming down and enslaving us. You can't rule it out. And if you think long and hard enough about the Drake equation, you do end up uneasy about it. Not to mention, you know, the Navy videos and all that. But at the end of the day, you're like, what can we actually do? And that is we can live and operate in history. And then again, I'm back to foreign conquest, environment, pandemics, which we kind of all agree are huge. And EA has done great work making those more focal. And I think they're things we can actually solve, not completely, but we can really make big progress on them in a way we can't with AI risk. If AI— if Skynet goes live happens, that's like a 3-sigma event, a 17-sigma event, and you just can't plan for those and you can't necessarily stop them.

So how do we prevent pandemics? Since, you know, that one's a little more achievable.

Well, there's so, so much we can do. I wouldn't say it's easy, but it's easy in the sense that we know what it is. It costs something. It doesn't cost 40% of GDP. You just need better FDA, CDC, better planning, better surveillance, better testing, da da da da. And there's so much medium hanging fruit there. I'm very focused on that personally, and I'm delighted to see how much of EA is.

I'm really excited about a bunch of this stuff happening in that space. I wonder if you want to tell us a little more about like the specific stuff you guys have funded, the— that you're most excited about.

Well, through Emergent Ventures, we made an early grant to a company called Curative, which repurposed itself toward doing testing and at its peak was 11 to 12% of all U.S. testing. They got their first money to buy the materials they needed from us. The spit test that came out of Yale, which is used for a lot of testing, it's actually better than the thing up the nose and people hate it less. We were the first people to fund that. Yale couldn't come up with the money to fund people working on that. That's insane. That's a super fixable problem. There should be like a fast grants thing at Yale that if you're doing something important, you can get the money for it in 3 days. And there wasn't. So like me, not my personal money, but me, the decision maker, that I ended up being the funder of that. Like, that's insane. That shows something is badly screwed up with the world. It's like if I ended up getting an Academy Award for being best actor, like the world has to be screwed up. Or if I won like MVP in the NBA, world's even more screwed up. Like I'm funding the Yale spit test. The world cannot make sense. I've seen so much low-hanging fruit. I wrote a policy paper, I think it was 2005, that we should prepare for the pandemic now. And it had some impact on what the Bush administration wrote into law. So I've been focused on this for a long period of time. So many things we ended up doing but might not have done, like re-gear up our domestic vaccine capacity. That almost didn't happen back in the Bush years. I was part of those discussions. It almost didn't happen. Now, we're fortunate it did. But again, that's just insane. Didn't cost that much. So that's one of the areas where we're just so far off base. And I don't think it matters. Like, are you a restrictionist, a Great Barrington, whatever? Like, you can debate those issues, but there's so much we should do that makes sense under virtually any view.

And I think when I get updates from fast grants about stuff they funded, there's the same sense of like You know, nobody else was funding pan-coronavirus vaccines. Nobody else was funding—

Yeah, we were the early funders of that also. And I hope those come to fruition. It seems they work. There's a question, does anyone care enough? Is there enough market that it makes sense to have the factory capacity? I'm not sure they will happen. I would bet like above 80% that they work maybe imperfectly, but still are a significant improvement. And we might not do them. That's— we're getting back to insanity again.

Man, I— so when they were giving out vaccines against monkeypox, I'm not in the at-risk group. I was kind of tempted to get it just because being immune against poxes and immune against coronaviruses, it seems like it'd be cool. I guess this is a very rare— we don't make up a market.

It's a very Bay Area attitude, perhaps.

So there's a lot of, like, frustrating institutional failures and inefficiencies. Are you still like, you know, pretty bullish on America? Is it the best place to make all this stuff happen?

Most stuff. I mean, I'm super bullish on North America and most of the Anglosphere. I've become a lot more pessimistic about many parts of the world in the last 10 years. I'm sorry to say, even the last like 3 or 4 years. But here there's opportunity, there's talent, the regulatory environment, even if totally screwed up, like you can get some things done. You have the English language here. It's critically important, major advantage, I think, still in a funny way underrated. But you— are you bullish on the US and North America?

I'm, you know, Bay Area. I see the more optimistic people. I'm like, yeah, we're building a lot of cool stuff. Like, it might kill us all, but I'm not actually very worried about long-term stagnation. I think we're going to build tons of cool stuff if it doesn't have any catastrophic effects.

I know a lot of Bay Area people, they end up super pessimistic. Because they see the social climate surrounding them and they overrate how much it matters. They're like, oh, the whole West is falling apart. We need to make some big intellectual move because of the wokies or something. And to me, they overreact. But that hasn't happened to you. Like, you've—

I talked to EAs and I talked to my neighbors who are like extremely chill, down-to-earth people who've lived in Oakland for 40 years and are like, wow, we have so much less crime than we used to.

So it's the Oakland thing. Yeah.

I brought this up because one of the listener questions is, what's your model for Germany being middle of the road in so many respects? I assume they're from Germany. It's a little bit of a rude question if they're not. Do you have thoughts? When they say middle of the road, lack of new companies and ambitious foreign policy, no strong university.

I think there's a lot of complacency in Germany. I think Germany has a problem fitting its ego into an Anglo view of the world. There's a longstanding Germanic allergy to Anglo dominance, which in a funny way, funny, unspoken way, has gotten worse. The whole German dream of living between Russia and United States is one of the worst mistakes like anyone in the world has made. Scrapping nuclear, actually part of that same mistake. A lot of infrastructure delays and problems. Primary education in Germany, like it's okay, but it should sort of be at the top. And it's not at all. In some ways it's getting worse. Vocational model— people have envied it for decades, but it looks a little stagnant maybe.

I don't know.

I'm— Germany is one of the places I'm less bullish on. I certainly think they'll muddle through, stumble through. But I think the German national psyche is still in a bad place. And there's not been a real, like, learning— just that, you know, they had Schroeder in power, basically bought out by Putin, no, like real hearings or coming to terms with that. So I'd say I'm worried about Germany quite a bit. And I used to live there. I feel I know Germany reasonably well. I don't know. I see more bad signs than good signs.

If you've, you know, lived in Europe and thinking a lot about Great Power War and the potential for evil actors, I'm sort of curious what lens you've seen the Russian invasion of Ukraine through, like should that be something EAs are thinking about? Should that be something the US is thinking about more?

Well, US thinks about it a lot. I don't know how much EAs think about it. I'm not sure what you can all do about it. I think it's one of the two most important events in the world right now, the other being China, Taiwan, Russia, Ukraine. With Emergent Ventures, just in the last week, we've been able to set up an Emergent Ventures Ukraine fund which will help Ukraine sort of restock itself with talent, give entrepreneurs there a chance to build or rebuild things. Frankly, I don't know how well that will go. Like, this is just a few days old. I am hopeful. I think it's super important. Ukraine has enough talent, but the invasion never surprised me. If you know history of that region dating back to the 18th century, Ukraine has hardly been independent. It's a mistake to pin it all on Putin. It's been a longstanding Russian tendency, even like Joseph Brodsky, the poet who, you know, emigrated to New York, lived there for many years. He, like, thought, you know, Russia needs to crush Ukraine. So it's deeply rooted. The problem is not easily solved. Think about it in a long-term point of view and you end up very nervous. It's been a flashpoint for a lot of world conflict. I don't have answers to it. I would say I'm quite sure I don't even know what current US policy should be. But I do think it's super important.

All right. It seems like the big thing our audience wants to hear now is things you dislike about the EA movement. We've got, what would you say are the most valid criticisms of the EA movement? We have, EAs care a lot about thinking clearly about uncertainty. Do you think we're good at it?

Well, this is my first EA event, so I can give you impressionistic senses of things, mostly from a distance. I'm not sure they're all criticisms. One, I think, is this. There's a risk, but also promise that you're all overwhelmed by your own demographic. So just to give you an example, I went to the Heretikon Peter Thiel event in Miami in January, and I was up here with Peter and there was a bunch of people in the audience who look exactly like all of you, sort of in the broad sense. I said, Peter, like, look at these people out here. I go to university events all the time. These people, those people, they look, they act exactly the same. Like, how are people at your event any different from the woke? And that was, in my view, this big embarrassing question with no really good answer. So there's educational polarization. There's like an immense amount of talent here. You're going to be successful on average, work a lot with other people. A lot will be diluted. That's part of how you'll spread your influence, but you'll be captured by that demographic. And there's a bunch of issues. Maybe social liberalism is the biggest one. Like, I tend to think EAs should be what I call voluntary extreme social conservatives. Live like Mormons. You don't have to be a Mormon. Like, I live like a Mormon. I don't drink alcohol. I could go on down the list. Most of you should live like Mormons. But my sense is you don't necessarily. And because demographically you're part of groups that are socially liberal, my sense is it's hard for EAs to kind of think straight about those issues. I had a dialogue with Will MacAskill. It's online. Will said, oh, the EAs, they're basically social liberals. I'm sort of like, where'd that come from? It's just demographics. So there's a tendency in EA to like talk through, reason through everything on these forums and sort of feel it's been digested and put it aside. But I would just say your demographic features will overwhelm the forums. Doesn't have to be a bad thing. I'm not saying it's a criticism. In large part, it's how you will be effective. But just to kind of internalize that as actual knowledge, I think you should all just be voluntary social conservatives and sort of turn your back on some features of your, your demographic. I think you should all be political independents. Neither party. I think they're both terrible. Like, on average, you'll end up being strongly Democratic because of your education achievement levels, coastal elite and all that. At least now I think that's bad, even though like I don't want, you know, Trump to be reelected. In that sense, I'm rooting for the Democrats. But there's some intellectual independence that will be very hard to square with the evolution of your own demographic. So that would be like my biggest observation. A related point is there's a kind of optimizing mentality, like eating how many chickens equal to the cow or the bed net versus the deworming. Like, say Edmund Burke were the best theorist of reality where he stresses like optimizing doesn't work. It's slow evolution, historicism. Like Edmund Burke is one of the best thinkers. How do you square the insights of Edmund Burke with how EA people actually think about problems? So I think you should all be more Burkean. But I again get like someone might hear this point, put up a point on an EA forum, you'll all talk through it, kind of decide the optimal level of Burkeanism. And go back to optimizing that, like, that's fine. It's still like so much better than where so many other people in groups are at. But it still seems to me a bit off, like actually just be more Burkean.

As you were saying that, I was mentally translating what you were saying into, I think, the like EA optimizing formulation of it. I was like, ah, yes, now I've translated what Tyler Cowen said into figuring out the optimal amount of Burkeanism to be.

But what's your biggest criticism? I want to hear from you.

Oh, my biggest—

I'd like to ask you.

Oh, I guess I think that too many people in EA, like, have figured out what the EA consensus is, but not figured out for themselves what they believe. This comes up a bunch with AI, like people who are like, all the smart people I know think AI is a really big problem, but like, I'm myself too confused. I think that's like just a terrible intellectual habit. Like, it is fine to think in the abstract because these maybe if I looked into it, I would agree with them. But like, you don't want to be in the habit of like having beliefs that are like kind of important to future planning and like priorities for the next 10 years that you're fully deferring on and that don't have any moving parts. I think people should like back off from anything where they're just, this is just what smarter people think. And also I think people are smart and can figure that stuff out for themselves and should actually just sit down and try and figure out what they believe. In a lot of cases where they're deferring. That would be like my big, big—

I would stress whatever differences I might have, like I grew up loving these kinds of ideas and the fact that there's so much talent in EA at this meeting, at the other meetings, to me that's far more important than the intellectual differences. So the economists focus like, what do people actually do? What you all actually did was to show up here. And I take that as a primary, like way more important than like how many pages of Edmund Burke are you going to read next week? So it's the same with crypto. Like, I know all the arguments against crypto. Maybe I'm a crypto agnostic or crypto hopeful, but at the end of the day, how many talented people have shown up to work on crypto makes me just a lot more optimistic about it.

Just on the principle that when smart, motivated people are working on something, there will be a lot of—

absolutely. And like so many of those people, maybe all of them know more about the thing than I do. Right. And they'll come up with something. And you want these clustered. It's kind of like Manhattan Project on altruism, kind of a mini Manhattan Project on crypto. We should do more like that. So you can't like be too picky about, oh, I don't like this Manhattan Project. Like, are you building— why don't you start with a hydrogen bomb? Or I don't know, like maybe they should have. But at the end of the day, it's like, let's do a Manhattan Project. Let's be excited about that. And I mean, that's my primary attitude here as well.

Right. So I got some people interested in if you do buy the premise that there's a non-trivial security risk from like power-seeking AI systems, is there like at all a response to that in like your state capacity libertarian perspective or like your perspective on the world? Or is it just like, I mean, if that happens, then that would suck. Like, are there things that if you were convinced of our beliefs there would, from your perspective, make sense to do.

I've never seen plausible alignment plans. Noting I'm far from an expert in this area, there could easily be a plausible plan I haven't encountered. The plans I do see, which are from smart people, I hear them described to me and I think, well, some evil person will just reverse the sign on that and you're actually an accelerationist. So I tend to fall in the, gee, if that happens, that would really suck category. We need to like go back in time and undo the whatever. Like they do in the Terminator movies. That would be my plan.

Why should we live like Mormons? Someone asks.

I think alcohol is very bad for people, maybe only 10 to 15% of people, but you don't always know in advance that that's you. There's social contagion from all drugs. So if just you don't do it and many other people don't do it, the people for whom it's a problem. It becomes far less focal. You'll be more productive. I think most people end up happier and sort of bourgeois lifestyles. I think on average are best for people, not everyone, but I think they're underrated and they're underrated by what you might call broadly like young intellectual coastal elites. So Because it's better for everyone and better for the rest of society. There's certainly a segment of the elite that can live an extreme version of social liberalism and then snap back to discipline when they need it. And I strongly suspect it's a lot of you out there. But if you think about society in general, I think we're learning it's not stable to have so many of the elite like that, and then the non-elite can't actually handle it as well. And that's a big problem. And one answer to that problem is just more people live like Mormons.

So, so live like Mormons concretely, like other than alcohol, is this like marry at a relatively young age, have a bunch of kids, live in the suburbs, attend church? You know, like which parts are you thinking of as important parts of— like, I think there's something here. I think that a lot of people in their early 20s are underestimating how much value, like personally and professionally, they get from like getting married and having a like stable, tight-knit, like social environment that's backing them. But I'm not sure which parts of that you're thinking of.

As some of that's context specific, some of that is change at the margin. I'm not sure we have social institutions where it's an equilibrium to marry early. Like the only people who might marry you early are the lemons. I get that. So it's not like, gee, You'd better go out and marry early. You like end up at the dumpster bin and, you know, hand over the ring. But I do think, you know, depopulation is a potential problem. Like someone has to have more kids. We can't just count on the country of Niger to be holding up the banner there. I think it's fine if more people like literally become Mormons, not just live like Mormons, but like, I don't believe in it. Like it's not going to be me. It's, oh, you, you, you know, you become the Mormon. And there's something kind of weird or even obnoxious about that. But still, it seems to me at the margin we should move in those directions, but with a lot of context and qualifications, of course.

Man. So I'm happily married. I have kids. I want to have 6 kids. I'm like partly there with you, but at the actual—

you're totally there with me.

At the actually become Mormon.

You leapfrogged me.

I'm like, no, it's really important if we're thinking about this weird future stuff to have true beliefs and be meticulously correct about stuff. You can't just become Mormon unless you become convinced that Mormonism is like making true characterization of reality and we become gods and get our own planet when we die, which would be cool. I like their afterlife schema. I don't think EA should do that if they don't think it's true. It's true.

But I think we observe, say, in Latin America, a lot of conversion of rural Latin Americans into Mormonism. And what they observe is just, it's good for your life, it means you're not an alcoholic, and they have some notion of the doctrine but not that much. So for the world as a whole, the sort of looser version of Mormonism, I think, is very much a live possibility. Here, becoming a Mormon is something much more specific, right? And like, say, in some rural Mexican villages, like literally half the men are alcoholics. And that you don't drink is the biggest change you make. And that is not true for current America. But still, always at the margin, we need a marginal revolution toward a bunch of features of Mormonism. And you go to the state of Utah, I'm sure most of you like, it's pretty awesome and it's better run. And I get the package deal and what if you're gay and not so great for women and All of that.

It's a pretty big all of that.

The point is there are ways at the margin that you can get a better version of the package, right? The whole point is to be a cultural, cultural entrepreneur who improves on the current packages but learns from the good things they do.

I guess the way I've always thought about this is like we had a, like, traditional set of norms and like it had some advantages, but it fundamentally doesn't survive exit rights. People, people don't stick with it once they have other options. People go to college once they have that option. And so whenever— and this isn't exactly what you're saying, but when people are like, we should go back, we should like try and re-adopt traditional norms, I'm like, they still don't survive exit rights. Like, the fuck is there going wrong? You have to like build something on the other side of like what we have now. You can't, can't go back.

I agree. We need a new version of them. I'm not sure the future is some inevitably totally secular. 40% of the people are polyamorous. No one goes to church. It's all an extension of what some parts of like England, Japan, or Sweden have done. I would just say, I don't know, like maybe it's impossible to even have a future version of these marginal changes, but I don't think we know. Our ability to predict that kind of thing is so weak. I would just say try to figure out what's good and let's shoot for it.

Checking what else people are—

and the giving up drinking thing like that's super easy. You can just do it. And I know a lot of people who've done it. They all like write to me or tell me how happy they are they've done it. Drinking rates are mostly going down, especially for educated Americans. They're going down. So that's not like, oh, can we really do this? Like, can we really talk people into 6 kids? Like, you can just stop drinking. That's why I start with that one.

I think that one is like quite reasonable. Like, the alcohol is very bad case, you know, even compared to other drugs. Like, you're not going to be as big a fan of this. It's not very Mormon. But if you just switch from alcohol to like other party drugs, I think you're usually making a good decision.

Could be. I don't even know, right? I assure you, I don't know.

Some of what you're saying seems related to one of the questions here, which is you've You've expressed skepticism that human life could be radically different in the future. And that feels like one of the differences between us here, because you're talking about the potential for the population to fall. And I'm like, yeah, I'm not very worried about that, because I think we're all going to be digital minds and make millions of copies of ourselves. And do you strongly expect that human life is going to look basically the same in 2100? There is so much potential for stuff to change.

I don't rule anything out, but I would go back to Edmund Burke and the historical nature of knowledge. Like the present is radically different from the past in a whole bunch of ways that mostly we could all describe. So those kinds of changes will continue, but like truly radically different that like we do symbiosis with an alien species, we're all uploads. I just think they're kind of knowledge, kinds of knowledge you can't theorize much about.. And the most useful thing to do is think of the future as different from the past in the same ways that our present is different from like 1680. And that's the best way to be productive about making the world better. And the other stuff, I'm not saying it won't happen. The aliens might come, the symbiosis might happen, Robin Hanson might be correct, we'll all be resurrected frozen heads or uploads. But at the end of the day, in terms of what you can do now, like back then, Environment, pandemics, war were the big problems. And today they're the big problems and we can make a lot of progress on them. Like, let's cooperate and do it.

I'm again, like mentally translating the things you're saying into optimizing things I could put on the forum. Yeah. And then the thing that comes out of that is something like you want to focus on the worlds where you have some steering power. Like if you think that there are lots of potential futures, but in some of them almost nothing you do matters, then it makes a lot more sense to like take actions that work, like actions that will affect the futures where you can absolutely.

Yes.

And I think I agree with that. And I'm just maybe slightly more optimistic that it's possible to affect things like the development of powerful AI systems than you. And maybe that's just like you're not huge on our institutions. You're not very confident we can do stuff that hasn't been done before.

We can affect AI in the sense we could accelerate it, we could make it better, make it do new and different things. I'm not sure we can stop the bad side very well, but I don't think we're without agency in the matter.

So what would like affecting AI in ways that you are optimistic about would look like, aside from accelerating it? You know, the most unpopular such proposal?

Well, as recent as 2 years ago, or even most people now, they'll push the Peter Thiel line like crypto is libertarian, AI is totalitarian. I get why Peter said that. You know, I've been to China plenty of times and AI now with facial recognition at Delta and so on. But we've seen these new innovations that appear to be very empowering of decentralized creativity and also criticism of regimes. Now we'll see how they evolve. But the people behind those innovations, I suspect they've given an enormous boost to decentralized human creativity in a good way. So like, let's do more of that. I'm less keen on surveillance. I strongly mixed feelings about it. I think it's probably going to happen anyway. It's certainly what I observe. And what Delta did is not controversial. Like, no one ever brings it up. And I take it seriously that no one has ever brought that up with me when I board Delta or when I show up in Heathrow. I'm just through the gate. And I think that stuff is inevitable. It could turn out to be bad. I'm not sure it will, but I don't think we're going to stop it.

So that I feel like I said, other than accelerating AI and then I got some like great AI stuff we can do, which is like building more and better.

That's what I think we can do.

All right. You've traveled a lot. How do you think your worldview might be different if you hadn't?

I think it would just be totally different if you travel a lot, especially to very different kinds of countries. You see different places as being more different. You have individuals as what I call regional thinkers. So influenced by their background. You see some universals to human nature, but you just get the sense of how much history matters and how much context matters, how hard it is to give general advice, how much you need to understand a lot of details about a situation. To give like one very concrete example, my colleague Brian Kaplan just wrote this piece on feminism. We were talking about this before. He calls himself an anti-feminist. Like, I don't fit. Fit in well with a lot of the feminism in the US. But if you've been to most of the world, to me it's just totally obvious, like the world as a whole needs way more feminism. Even if you think like some of this stuff in whatever direction goes too far, and you should not at all take an anti-feminist stance. And that comes, I think, you know, from traveling. It's just obvious to me. Not long ago I was in India, I was in Pakistan, and so on. Very different perspective.

Do you think that most EAs, you know, as part of getting out of the coastal demographic bubble that you warn about, should travel more? If so, like where? Because I totally think of it as like a coastal thing to travel and then continue being the kind of person that you are.

I don't know how much you all travel, but I think most people should travel more. And if I meet someone who's been to 40 countries, I think they should travel more. And on average, go to weirder places or places you haven't been if you can. Like, you may not be able to, but I think it's by far the best form of learning. Way better than books. Books are overrated. Not at the beginning. Like, the Quake books are amazing and immense, and you sort of need them as a foundation, if only to help you understand your travel. But at the margin, you've read like all the books everyone else has read. You ought to travel. You can read while you travel, read on the plane. Like what? Don't watch the crummy movie. The screen's terrible. Like, that's one of the biggest inefficiencies of human life is how many people stare at screens on the plane, which is a way worse experience when they should be reading. And reading on the plane is really not a worse experience. And yet everyone's staring at the screen. And that's true. Like, even in elite business class, first class, it doesn't matter, like how smart the flight is. Like, ooh, New York to Paris. These are the smart people. No, they're staring at the screen like, read the Book of Mormon, if nothing else.

I have a bunch of overrated or underrated questions here. Mormons are underrated.

Well, not by Mormons.

The United States, underrated or overrated?

Way underrated. Now, it's underrated for different reasons. I think people in America who are kind of the coastal intellectual elite, underrate it because they think its politics is totally falling apart or it's become like so polarized or so evil or so screwed up. Like, there are some concerns which I share, but people take that way too far. A lot of foreign countries, like maybe we're overrated in like Ghana, Nigeria, Honduras, a few places. But there's so many people kind of tricked by our own media and they think the US is in much worse shape than it is. And if they come here, they're like, kind of seems the same or it seems pretty normal. So they're underrating it too.

EU underrated or overrated?

It's way underrated by Americans. It's been this fantastic thing with free immigration, semi-free trade, has kept the peace. It's probably overrated by a lot of people in France. Maybe properly rated by a lot of Eastern Europeans who properly understand it as their lifeline for liberty. So it depends where you go.

Plant-based meat, overrated or underrated?

I've always been suspicious. I think people are biologically programmed not to mind eating real meat and to want to eat real meat, even if the other stuff tastes pretty good. So like, I wouldn't buy shares in those companies. So maybe it's way underrated by me. But the fact that the ordinary American is not that interested shows, like, amongst those people, it's probably properly rated. I get it would be good for the world. I have no disagreement with those arguments, but I just wouldn't bet on it.

Progress Studies, overrated or underrated?

I don't know how it's rated. Like, I'm not even sure what it is. It's not a movement like EA is. It's a set of ideas that try to have influence without being embodied in a movement. So in that sense, you could think of it as a different strategy.

Is this strategy you think will work?

I think it's— forgive the tautological self-invocation— it's making a lot of progress. So just a simple example in my field, economics. Economics of science has been a tiny thing forever. That makes no sense to me. It should be at least 10% of the field. In the last few years, I've seen it grow considerably. Now, most of the people doing it wouldn't call it progress studies, or I don't even know if they've all heard of it. But like, it's happening. And I think that will happen in many more areas. What's like the final translation into results? Harder to say. But I just see a lot more interest in things like yin-bi-ism or immigration to boost growth or history of science, practice of science, a lot more interest a lot of it driven by EA people, it seems also. But I don't have some grand prediction for it.

Xi Jinping, overrated, underrated?

Well, in what sense? I think he's an evil guy. I think outsiders who inform themselves about China probably rate him correctly. When I used to go to China pre-pandemic, people actually really liked him. It wasn't just they had to say it. I don't know if that's changed with zero-COVID policies, but he was then overrated.

He was then overrated by the—

He was genuinely popular in China less than 5 years ago. How that has developed now, I couldn't tell you. Risks of nuclear war, probably by this crowd properly rated, but I think underrated in general. That's what can do real damage. And there's just recency bias. We haven't dropped bombs since Nagasaki, you know, against people. And that means we forget about it. It's not a thing. When I was growing up, fear of nuclear war was a major thing. You know, there would be drills and, you know, signs up and everyone talked about it. I was born in 1962. Cuban Missile Crisis is a year later. We all thought about it a lot. And it seems to me that's receded in a way that is very bad.

It's interesting, like watching old movies, how often nuclear war is the apocalypse that they're preventing.

And these days it very rarely is, even though, you know, they rated it properly. It's a little tricky for EA people. It's one of these, well, what can I do about it? Very hard to say. In a way, with pandemics, it's not hard to say. It would be weird if the right answer was zero, but it's not that simple. With pandemics, you'd have a list of 30 things off the top of your head.

Mrna vaccination as like a potential future tech.

I think biomedical advances of the next 30 years are still underrated, but mRNA probably won't be the most important thing. So amongst people in the know, maybe slightly overrated, underrated by the general public. Biomedical advances in general, way underrated by almost everyone. I think the next 30 or 40 years will be this incredible golden age. Where we'll fix so many different things.

What are you most excited about in that space?

Kind of everything. Like, I don't think people will live forever, or maybe we can't pass 120, 130, but big progress against cancer. The malaria thing already seems to be working. A lot of progress on dengue, sickle cell anemia. I just think most things that are major problems, we're going to see immense, like more than 50% progress within the next 30 years.

Do you think that the EA movement should be like trying to do more? You were talking a little bit about how we ask like, what should we do individually? Should we be doing more like to go into government and policy and like change big things instead of like do stuff that you can do on the margin?

I don't know what you all do now, so I'm not commenting on you all, but I think smart people in general underinvest in going into government. It pays less. It's more frustrating. And to have impact, it has to be done collectively. So that's in general true. It's probably true for EA people, but I don't actually know that.

What would— like, what developments in EA would you be like most excited about? Like if you came back here next year and we'd done a bunch of stuff, what stuff would make you feel like most optimistic about where the movement was going?

Less moral nervousness and more focus on what is my passion and building outwards from that, I think. Would make me happy. But like, you know, just keeping on with pandemic, which seems to be going great, that not changing.

Are there things you might see in like the upcoming years from AI that would make you rethink some of your thinking about AI? Like if you saw an alignment solution that seemed like it might work, or if you saw some like really destructive products of further AI research?

If it kills me. But no, seriously, I think the real issue will be using non-sentient AI to fight highly destructive wars. And I think I will see that. So it's not hypothetical. I just don't know how bad it will be. And the real question could end up being like, who is in fact the bad guy? You can never take that fully for granted. Who's the bad guy.

So you think you'll see that, but you don't think there's like much to be done on the development side to like reduce the odds of that or like slow things down so that doesn't happen until we have a better plan?

I think the Pentagon and China are in a rivalry. They're both optimizing locally and there's no feasible way to break them out of that equilibrium. You might even think it's prisoner's dilemma. I'm not sure it is, but it certainly might be right. Like they'd both be better if they stopped. But I don't see how you get there. It's like, how do we get to families where they all have 6 kids? Whether you think that's good or bad, I would just say, present company aside, perhaps I don't see how we get there. Like, it's not on the table. So, no, I don't, don't think we're going to square that circle. All right.

What are the biggest economic challenges for the US or the world right now?

Biggest economic challenges?

Yeah.

The overall social and mental health performance of our bottom third. Is by far our biggest challenge, and I'm not sure what to do about it. I think it's a great topic for EAs if you have the passion, but it's much harder, say, than pandemic. What do you do to improve mental health in this and other countries? To me, very slippery topics. A lot of stuff doesn't work that well. Better policy. No one's against that, but I'm not sure it's really going to make that much headway or gain that much traction.

All right. We're a little bit low on time and have a lot of votes for where should we go for dinner tonight? So do you have thoughts on that?

Depends on the size of your group. Lately in D.C., I've been enjoying Rumi's Kitchen, which is Persian. It's on— it's not close to here, but it's not too far from here. There's a Mexican place called Destino, and there's a taquería too. Side, the same management, which I think is excellent and has good outside dining. They're the two D.C. places I'm fond of lately. Best food, we all know, is in the suburbs. Sorry, people, you're trapped here. But if you take an Uber to Arlington, which is really not far, it's closer than a lot of D.C., Peter Chang Sichuan in Arlington is excellent and gives you like a look at real life, not just Washington life. I'd recommend that too.

One more overrated, underrated. The Biden administration.

Well, it's either one or the other, right? It depends who you are. I would say on COVID, I am deeply disappointed at how bad they have been. Like Biden was good on the boosters when he didn't have to be and faced a lot of pressure pushing back against him. Very happy about that. But almost everything else, they've been bad. No generalization of the Operation Warp Speed model. The reform of the NIH, I guess I have some hope for. That's a good sign from them. That's very recent. And then it's just issue by issue. And on all that, like, we just have to try to be objective. It's very hard to be like no one really is.

Yeah. All right. That's about the end of our session. But if you have more questions for Tyler, which many of you clearly do, he's doing an office hours just across the hall in Oceana. So you can head over there and and continue asking questions. Thank you so much, Tyler. This is—

Thank you, Kelsey. Great to meet you.