Podcast Episode Summary
The podcast episode features an interview with Dr. David Friedman, an anarcho-capitalist economist and author of "Legal Systems Very Different from Ours." The conversation covers a wide range of topics, including the efficiency of the dating market, the enforcement of contracts in a world with strong privacy, and the potential consequences of a reputation-based system.
Main Topic: Efficiency of the Dating Market
Dr. Friedman discusses the dating market and its inefficiencies, citing the difficulty of finding a compatible partner due to differences in tastes and preferences. He argues that online dating has not improved the situation, and that the problem is more complex than a simple sorting issue.
Key Points:
Participants:
Takeaways:
Additional Topics:
Final Question:
Dr. Friedman is asked what field he would decide to go into if he were 20 years old, and he suggests independent computer programming, citing his enjoyment of computer programming and his ideas for creating clever programs to help teach economics.
David Friedman is a famous anarcho-capitalist economist and legal scholar.
Watch on YouTube. Listen on Apple Podcasts, Spotify, or any other podcast platform.
Episode website + transcript here.
David Friedman's website: http://www.daviddfriedman.com/
Follow me on Twitter for updates on future episodes.
Timestamps:
(0:00:00) - Dating market
(0:12:15) - The future of reputation
(0:27:30) - How Friedman predicted bitcoin
(0:35:35) - Prediction markets
(0:40:00) - Can regulation stop progress globally?
(0:45:50) - Lack of diversity in modern legal systems
(0:54:20) - Friedman's theory of property rights
(1:01:50) - Charles Murray's scheme to fight regulations
(1:06:25) -Property rights of the poor
(1:09:07) - Automation
(1:16:00) - Economics of medieval reenactment
(1:19:00) - Advice for futurist young people
Dwarkesh Patel
Today I have the pleasure of interviewing Dr. David Friedman, who is a famous anarcho-capitalist economist and most recently the author of *Legal Systems Very Different from Ours*. So my first question is this: is the dating market efficient in the sense that it doesn't maximize net value? It's not a competitive market because I'm not substitutable for another bearded 20-year-old Indian man. So I'm curious how efficient you think the dating market is.
David Friedman
Oh boy. I've been puzzled for quite a while on why online dating doesn't work better than it does. In principle, a lot of dating is a sorting problem. Not only are you not identical to another bearded Indian man of the same age, your potential dates have different tastes. I actually discussed some of this. I've got a chapter in my price theory text, which is really on what I think of as the marriage market, although it applies to dating more generally.
I have one model, which is the really simple economic model, where in effect everybody is identical. You have a market with an implicit price in terms of the terms of marriage—how much the husband agrees to do, to what extent the wife gets to make decisions, things like that. You then allocate people and have an equilibrium price. The real point of that section is that I'm arguing the effect of polygyny or polyandry is the opposite of what people expect. If you legalize polygyny, allow a man to have multiple wives, the result is to benefit women and maybe harm men. Because if men can have multiple wives, that bids up the price of a wife. Some men can go on the market and try to bid for two wives instead of one. That means the ordinary monogamous man still ends up getting a wife, but on worse terms—he's got to do more of the dishes. Women either get all of a husband on better terms, or half of a husband on terms good enough so they prefer half of that husband to all of another. That's a fun argument. And of course, if you legalize polyandry, allow a woman to marry multiple husbands, you have the same effect in reverse.
So that's the easy model. Then the harder model is the sorting model. Even then, I only have a simple sorting model. My sorting model is one where everybody's value as a spouse varies, but it's the same to all potential partners. The real one is where not only is your value as a spouse different from somebody else's, your value as a spouse is different to some people than to others. The way I like to put it: the woman who I eventually concluded was about a one in 10,000 catch was not being pursued by anybody else at the time because other people had weird tastes and did not recognize what a desirable partner this was from my standpoint. I should say we're still married. That was 38 years ago or so that we got married.
So that's a really hard problem. I'm not quite sure. A lot of economics ignore search costs because we don't have a very good way of handling them. Yet search costs are really central to the dating problem. I would have thought things would get a whole lot better with computer dating because you can take all the characteristics that are sort of objective and measurable, tell them to the computer, have each person search a million profiles, find the subset that meet the easy characteristics, and then interact with each other to find how well they do. But as far as I can tell, it hasn't happened. I don't think marriage has worked better since computers came in. The people I know who have been trying to use computer dating haven't found it very successful. In fact, the story I'm told is that OKCupid, which is one of the most successful of these things, has gotten worse over time. It was originally designed by enthusiasts who did clever things, then taken over by businessmen who did less clever things. It may be the people I've talked to are just sufficiently different from the mass of the market so that what works for other people doesn't work for them. That's my impression.
So is it efficient? You've got to compare it to some alternative. It's certainly inefficient compared to the ideal of a world where a perfectly wise, benevolent person paired us all up, but we don't have any of those people around. The interesting question is: does it work better than the system where parents choose your spouse for you? Quite a long time ago, I had one of the most interesting conversations I've had. I was in Bombay waiting for a plane to Sydney and got into a conversation with a woman from southern India. We ended up sitting next to each other and talking a good deal all the way to Sydney. She was flying out to join her husband, who was a physician. Her husband had been chosen for her by her parents. She would have had a veto—it required her consent—but the actual search was done by her parents, not by her. She was still happily married. My first marriage had broken up. So the evidence from our little tiny sample was in favor of her system.
What was really fun about it was that here was an intelligent, modern person, an educated contemporary from an entirely different society in terms of that set of institutions. It's a lot harder to say, "Well, obviously, the way we do it is right," when you've actually got someone real there defending the alternative. So I really don't know whether the modern American system, the modern Western system of courtship, works better or worse than the traditional system where the parents find a potential partner, then the partners get to decide whether they approve of each other or not. There is the argument that there's hardly anything farther from the economist model of rational choice than a man in love.
Dwarkesh Patel
Interesting. I wonder if, to the extent that arranged marriages sometimes work better than non-arranged marriages, I wonder if that's because of a social structure that maybe makes you happier while you're in marriage, versus whether your parents are better at choosing a mate for you.
David Friedman
What do you mean? You mean it makes you happy because you don't feel as though it's your fault, that you messed up and should have done better? Or maybe those cultures are just better at making wives and husbands stay together. In a sense, what you'd really want to look at—and what exists, though I don't want to share their data on it—would be people from India, particularly Southern India, who live in the US. My impression is it's not all that uncommon for a young adult, maybe in the tech industry, to migrate to the US, then go back to India to marry the wife his parents have chosen for him, or in some cases arrangements are made here, and then live in America. That's as close as you can come to a controlled experiment, separating out quality of choice versus the degree to which the culture supports your marriage.
The other way you could do it, which would be much harder, is what you'd like the computer dating people to do, but I don't think they've really done: serious research on what observable characteristics match with what. In principle you could do it. You'd like to say, "Alright, here are all the things somebody tells us about himself, because that's really all you've got in practice. Let's take all pairs who ended up married, or maybe all pairs who dated, who made the first try of having a date. How do you figure out from the characteristics of the two parties whether they'll have a second date?" You would think they'd want to do that. Then ideally, you follow that through and say, "Which ones got married? Which marriages lasted?" So ideally, you could get statistical data saying, "Here's how to pair people up." If you had that, you could look at couples produced by American courtship institutions and couples paired up by their parents and see which one came closer to matching what the computer told you you ought to get together. That would be a really neat project, but it would be very hard to do.
Dwarkesh Patel
Yeah, that would be interesting. In *Hidden Order*, when you talk about efficient markets, you explain it in terms of: could a bureaucrat god improve on the market by changing the allocation or production or quantity produced? In this case, it's funny. In a sense, the parents are that bureaucratic god. Except the parents in those societies have almost enough authority—they could be rejected, but usually aren't. But they aren't perfectly wise.
David Friedman
Yes. The bureaucrat god is my attempt to make sense of what economists mean by economically efficient. There are two issues with regard to economically efficient. One is what's an improvement. There, as you know, I go with Alfred Marshall's view rather than the more conventional Pareto, Hicks, Kaldor approach, which I regard as a way of pretending not to make interpersonal comparisons when you really are doing it. Once you've got your way of comparing, when you say this is efficient, efficient compared to what? As far as I can tell, the closest I can come is to say: what's the best outcome you could imagine by this criterion?
That's my bureaucrat god—an omniscient, omnipotent, benevolent actor controlling everything. Of course, nothing will live up to that, but some things will come close. The standard price theory efficiency theorems tell you that a competitive market without certain specific problems, such as public good problems and externalities, would actually produce the efficient outcome. You never quite have that, but it's a whole lot closer than if you try to imagine other institutions, such as a centrally planned system, where the central planner has neither the right incentives nor the information nor the power that the competitive market has.
The market has all the power you want because each person controls himself. That's control over everybody, just in decentralized form. A central planner never has that.
So I think it's a useful concept, but you have to recognize that once you get to a fully realistic picture of any system, you're not going to achieve it. All you can do is say, "In this approximation, you would achieve it." Then you have some feel for how much breaking the approximation messes it up. It's a useful way of thinking about economics, in my view.
Dwarkesh Patel
I want to ask you about futurism now. You talk in your book *Future Imperfect* about how contracts will be enforced in the future. And you see, because of the rise of—
David Friedman
They will be enforced in the future. One of the things I try to be careful about in that book is to say I'm not making predictions. I am saying: if this technology develops in this way, this is what will happen. But I don't know which technologies are going to develop in which way. In the case of contracts, I think you're thinking about my discussion of how a world online works, where you have complete privacy—what I refer to as a world of strong privacy. Encryption could potentially give you a world where I can do business with you and no third party can observe what we're doing, including making money. The question is: in that world, how do you enforce contracts? I say, if that happens, if you end up with that future, here is a plausible way contracts could get enforced using ultimately reputational mechanisms.
But I don't know that world is going to happen. The National Security Agency has been doing its best for decades to keep that world from happening. As I try to make clear in the book, there are downsides of that world as well as upsides. We're seeing one of the downsides already. One of the things I think I discussed in *Future Imperfect* is the possibility for cyber ransoming—getting onto somebody's computer, encrypting the whole contents, and then saying, "I'll sell you the key back to decrypt it." That's a bad thing. You can think of other bad things. If you kidnap people in the real world, the hardest part of being a successful kidnapper is how you collect the money, because that's where you're likely to get caught. We can solve that problem with good cryptocurrency and a world where you've got enough encryption so people can't follow who's sending messages to whom. That world has downsides as well as upsides.
I happen to think that governments do more harm than good, to put it simply. A world where governments find it harder to control people, although it will have the disadvantage that some of the things governments do are good and won't be done, will have greater advantages. But in any case, I'm not predicting that world will happen. I'm saying that world could happen. There's the technological possibility for that that we've known for decades at this point. Bitcoin is newer than that, but cryptocurrencies are the only new part of that world. I wrote this up long before cryptocurrencies existed. So we could have that world. It's a really interesting world, which is part of why I talk about things like how you enforce contracts in it. But I'm not predicting it's going to happen.
Similarly, for my various other things I discussed, the one I most want is to solve the aging problem. At the moment, under current circumstances, the probability that I will be alive 30 years from now is very close to zero. I would like to be alive 30 years from now. Living is a lot of fun. The particular future I'm most in favor of is the one where they solve the aging problem, where I get at worst to stay at something like my present age—a little bit creakier than I used to be, but tolerable—and ideally get to run it back a couple of decades. That world will have its own features. Mostly good, but not all good.
If that had happened a few decades earlier, it would have been Stalin who presided with nuclear weapons over the breakup of the Soviet Union, and he might have had a different view of the people in power. Spain would still be ruled by Franco. So there would be bad things about a world where people didn't get old, but I think on net there would be good things. All of these futures have positive and negative features, and I'm not predicting which ones will happen. I was only trying to explore each of those futures. There's a reason I called the book *Future Imperfect*. That was partly a grammatical pun and partly the fact that none of these futures is perfect.
Dwarkesh Patel
One of the early cypherpunks, who were exploring the encryption stuff before anybody else, his line was: "Encryption is not your friend." Technologies are not your friend. Technologies are what they are.
David Friedman
Yeah.
Dwarkesh Patel
So, okay, fair enough. Assuming we move to a system where more contracts are enforced by reputation rather than the traditional legal system, I want to ask you: what are the social consequences of that? For example, will we see a return to honor culture? And is the rise of political correctness maybe a consequence of us turning to a culture that depends more on reputation? Because now whether you are woke or not says something about your reputation, and that signal matters more than it did in the past to your employer or to potential clients.
David Friedman
That's interesting. The obvious problem is that for some people, woke is a positive signal and for some a negative signal. Certainly for me, it's a negative signal. The thing is, the reputation you want is an online reputation, which does not have to be linked to your real space identity. In the kind of world I was discussing, it wouldn't be linked to your real space identity. Partly it's the fact that it's not linked that means you have to depend on reputational mechanisms. But of course, reputational enforcement has been important for a very long time. I would guess that most of what makes stores behave themselves, act well, is not the legal threat at all. It's the threat of bad reputation.
I had the experience of having problems with the people putting in the battery for my solar system. I put up a very negative review online and immediately got a response from them: "What can I do to help things?" They ended up giving me a $500 discount to apologize for the month during which they had turned off the solar system that already existed when they installed the battery. There are a lot of other examples nowadays that don't even depend on transactions being online. Companies care about their reputation, and they have an online reputation even for real space transactions.
But the question is: is there any particular reason why you would trust somebody who agreed with your politics more? My impression is that woke people quite often fight with each other and attack each other, try to cancel each other because they're the wrong version of woke. So you might get the pattern, which you already have, where you've got groups of people who have something in common. One of the odd patterns you get is some particular market niche, say donut shops, being dominated by some particular ethnic group. I think donut shops are a real example, but I don't remember which ethnicity. From my own experience, running motels is dominated by Indians in the US. I don't know why, but that's an observed pattern. Part of what may be happening is that you have an ethnic group. You don't want to offend somebody whose son might want to court your daughter. There are a bunch of social links within the ethnic group. Even aside from that, you don't want to offend somebody who knows people who know you.
I remember observing a long time ago online, back in the days of Usenet, which was really before the web. One of my hobbies is a group that does historical recreation called the Society for Creative Anachronism. It had a Usenet group, and it was a whole lot more polite than most Usenet groups I observed. I think the reason was that if you behaved really badly on there—went around insulting people for no good reason, being unreasonable—even if the person you were being nasty to would never meet you, he probably knew somebody who would meet you. You didn't want to have your real space reputation messed up.
I think that's an effect which will become smaller in the future world, because in the future world we've got an online version of that. In the future world, at least as I've described it, I don't have to know you in real space. Most people I'm interacting with, I won't. Many won't even be on the same continent. I've interacted with a number of dealers in India who I've never met, but who I'm willing to trust to deliver what they say they'll deliver because they've done so in the past. It seems to me that the internet means reputation is less tied to real space interactions and more to online interactions.
We have better mechanisms for keeping track of reputation than we used to. On eBay, for example, when you have a transaction, you get the opportunity to report to eBay on how it went. When you're considering dealing with someone, one of the things you see is their eBay reputation—how many people have said they fulfilled what they said they would. You get the same thing if you buy something from Amazon that is not actually being sold by Amazon, because Amazon also funds the middleman. Books are the context where I'm most familiar with this. You see: this seller has it at this price, and here is what his rating is, what people who have dealt with him say.
So I don't think I would expect—in a sense, I think of what you're calling honor culture as more the real space version of doing this.
Dwarkesh Patel
Interesting. I wonder what the consequences are of tearing down the walls between communities that perhaps make reputation in some ways more salient. For example, with the Indian motels: I've lived in small towns across the US. You might have at most 10 people of Indian origin, which isn't a big deal for me, but for my parents, that's a very big part of their social community. Losing status in that community is a big deal, and reputation matters a lot more. Similarly for your Usenet group for the SCA. But when you have communities that aren't isolated in that way—for example, a Twitter user—is the value of reputation, or the capacity of reputation to tamp down bad behavior, reduced?
David Friedman
It depends a whole lot on whether you are interacting with people where your reputation costs you anything. If I become part of the Twitter mob online, insofar as I have an online identity, that online identity is not doing business with anybody. It's not involved in any interactions where people trusting it matters. So it's free to behave as badly as it wants. On the other hand, if I am interacting on a blog or a forum with maybe a hundred people who are regular posters, reputation does matter. If I behave in a way that shows me to be an unreasonable person, people won't take what I say very seriously. It's not as serious as "he will not be willing to sell to me or buy from me," but it's still a real cost.
How this is going to pan out—the internet is having and is going to have non-obvious effects. I'm reminded: I just saw somebody online who quoted a bit that Heinlein wrote, talking about the influence of automobiles. He said there were people who, when the early automobiles came out, suggested this would end up entirely replacing the horse. But he doesn't know anybody who realized how it would affect courtship behavior. The existence of closed automobiles meant there was now a place where couples could make out where their parents wouldn't be seeing them, and where they could be 15 miles away from anybody who knew them, reducing the chance of reputational costs.
Somebody else online quoted somebody 30 years before Heinlein, when this was just happening, who was commenting on this effect. A judge in some rural area said that of the last 30 women who had come before him charged with sexual offenses—which I presume just means having sex, though he didn't clarify—19 of them had committed them in an automobile. That was an interesting case where you have a technological change with some obvious effects, but some less obvious effects. Similarly here, I don't know what the final results are going to be. I don't know how much the whole woke Twitter mob cancellation phenomenon is only happening because of the internet, or how much it's due to other ideological changes over time.
It's very hard to know.
Dwarkesh Patel
Speaking of unexpected developments, you have a chapter in *Future Imperfect* on e-cash. The book was published in 2008, which is the same year Satoshi Nakamoto wrote his white paper on Bitcoin. In many ways, it was prophetic, the chapter you wrote. But in some interesting ways—and this is in the context of you being very right—what Bitcoin actually was was a completely different paradigm than what you imagined e-cash being. Specifically, you imagined it would be some centralized institution like a private bank or company instituting money underwritten by either dollar denominations or commodity bundles. But in fact, you have a completely decentralized currency. I'm curious: in retrospect, what do you think you missed? How do you think you got it so right?
David Friedman
Well, I couldn't have predicted Bitcoin. Nakamoto was a very smart guy. I've done computer programming and kept an eye on encryption issues because they interested me, but I was never at that level of expertise to realize what he could do. It was a very, very clever idea.
The version I was thinking of was Chaumian digital cash, the math of which had been worked out by Chaum maybe a couple of decades before I was writing something like that. In a sense, that's the biggest difference. Because Chaumian digital cash requires an issuer, it hasn't happened. You need an issuer which is a reliable institution, which means it almost has to be in some developed country with reasonably well-functioning legal institutions. But one effect of anonymous digital cash is that money laundering laws become unenforceable. Governments very much dislike that, in part for good reasons, in part for bad reasons. Consequently, no country of the sort you would need existed that was willing to put up with it. That's at least my interpretation of why it never happened.
The beautiful thing about Bitcoin is there's no issuer and it doesn't require that. It's just happening invisibly online, which is very clever. The problem with Bitcoin, but not with cryptocurrency in general necessarily, is that the value of a Bitcoin is very unstable over time. That's not too much of a problem if you want it as a speculative investment. I in fact have some Bitcoin. But it makes it rather inconvenient if you're using it for day-to-day transactions. If a price is set in Bitcoin, the seller doesn't know what value he's going to get for his good. The buyer looking at the price in Bitcoin can't easily compare it to what somebody else was selling the same good for a month earlier, because the price could have changed 30% over that period. So Bitcoin is really not very good as money for transactions.
That means you want somebody to produce a stable coin that's easy to use. There are various projects of that sort. I've got a blog post in which I sketch out one mechanism by which you could get a stable coin, stable in dollars or something else. One thing that ought to be doable, but I'm not enough of an expert to know quite how, is to have a Bitcoin where every time the value goes up, you get a new subdivision with the same value. So in effect, if the value goes up, instead of having 100 units, you're now having 112 units. If it goes down, instead of 100 units, you have 90 units. Then the units themselves could be fixed to the dollar.
I tend to write more than I read, so I haven't really made a serious effort to keep track of the various projects for stable coins other than the one I invented, which I discussed on my blog. But I think if you had a cryptocurrency that had a reasonably stable value in terms of dollars or euros or Swiss francs—something whose exchange rate to other things didn't change very rapidly, which is what you want—and if it was easy to make transactions in it, which I think is not yet true of Bitcoin, then it would become a serious alternative to conventional currencies.
If you could do it in a way which proofed it against inflation, meaning you don't lock it to the dollar but maybe lock it to the average of the dollar, the Swiss franc, and the euro, then the basic problem is you need an oracle. You need a way in which the software can find out what the exchange rate is. I know of some clever ways that some of the people with the second most popular currency have interesting ideas about that, because I discussed it with them a few years ago. But I don't really know to what extent people have oracles that they can trust.
The more the software can determine facts of the real world, the more neat things you can do with it. There's this old idea of smart contracts where you make a bet with someone and the bet is in effect programmed into the money. If I win the bet, you suddenly have less money and I have more, and vice versa. There are ways of doing that, but it requires that the software can find out who won the bet. That's an interesting project which requires ingenuity to solve.
I should say my ideas on cryptocurrency are partly because I was interacting with some of the cypherpunks very early. One of them had stolen a bunch of ideas from me, reprocessed them, and I then stole back his version. I had the discussion in *Machinery of Freedom* of how a stateless market society would work, and he had said, "Well, what happens with those ideas if you have strong encryption?" and had worked out some neat stuff. Then I read his stuff and borrowed it back. So I can probably claim a little bit of credit, but it's not like I thought up these ideas of digital cash all by myself. I was partly part of people who were interacting over that. I wasn't really a part of the cypherpunks. I was on the mailing list for a while and I met some of them once. But I knew what they were talking about and it was neat stuff.
Dwarkesh Patel
I've done a little bit of programming with smart contracts. One of the very interesting ideas is that you can have contracts that own assets. The contract owns Ethereum, which in retrospect seems like an obvious idea, but it just seemed like a very cool idea. Speaking of smart contracts, I'm curious what you think about prediction markets.
David Friedman
I have known about prediction markets for longer than almost anybody else because I was corresponding with Robin Hanson back before he switched fields and became an economist. That's my standard example of one of the really nice things about this world. It used to be that you had to be a university professor because the people you wanted to interact with had to be in the same building. Now it's no longer true. I came across Robin very early, discovered that he had these really neat ideas, and so I corresponded with him.
Prediction markets are a very interesting idea. I gather they are developing. The ones I know about seem to have limits in the size of the bets you can make and in the amount of house cut if you want to take your money out. I assume it will develop further. Robin's idea, of course, is that you want the prediction market as a way of generating information. A lot of great ideas are very simple. His idea was that a stock market has as its nominal function allocating capital, and just as a side effect it generates information because you observe that certain stocks are going up or down, which tells you something about what people who are willing to bet on it believe about those companies or industries. So he said, "How about you do it deliberately? There are things we'd like to know. What if you set up a market for that purpose?" That was where prediction markets originated, which was Robin Hanson's idea.
Some have always existed. If you think about futures markets for agricultural products, a futures contract is a prediction market, although they're not referred to that way. If you observe that the future price of wheat is sizably larger than the present price as of a year out, you say, "Oops, looks as though the smart money expects that there is going to be a bad harvest or something else is going to happen." That's very neat.
In principle, it requires that there is a reason why people want to bet on this, but people like to bet on lots of things. I think that's something nice that's developing. I haven't really paid enough attention. I haven't actually ever made a prediction market bet, although my younger son has been betting on PredictIt and says he's ahead so far. He's doing it mostly for fun. But you don't want to limit people doing it for fun, because if you can only make small bets, then the prices are going to be dominated by ideologues. Whereas if you can make bets of any size, then the smart money who doesn't care what's true, they only care what bets make money, is likely to take over.
There was an interesting case back during one of the US presidential elections a few presidents ago. The relevant market very briefly surged in favor of one of the candidates and then very quickly went back. My conjecture was that somebody who wanted that candidate to win thought that pushing the prediction market in his favor would make people expect him to win, which would then make him more likely to win because you're more likely to donate money to people who you think are going to win. But that effect got wiped out by the smart money coming in on the other side when they saw that there was free money lying around. That does require a market where you can make reasonably large bets, because otherwise it's the number of bettors and not how strongly they feel about it that matters. Prediction markets are a neat idea.
Dwarkesh Patel
In the chapter on e-cash, you talk about how at the time you thought the burdens to e-cash were regulatory, not technological. When we talk about these other fields, to what extent do you think progress is stopped by not having the right technology versus regulations not being conducive to innovation?
David Friedman
That was my view then, and I think I may have been too optimistic. It's very hard for regulation to stop progress because there are many independent countries. The evidence against that is that as far as I know, nobody did challenge trials for COVID vaccines. Nuclear fusion may just be hard, but we know vaccines aren't hard because they produced a successful vaccine in something less than a week. People talk about how wonderful it is that we got vaccines in a little under a year, but the truth is we got vaccines in less than a week. All the rest of the time was the testing procedure.
You could have done a good enough testing so that it should have been available to anybody who wanted to take the risks in maybe two months. What do you do? You do two tests simultaneously. One is you give the vaccine to a large random set of people—old, young, et cetera—and you don't expose them to the disease. All you're trying to do is see if the vaccine has serious side effects. You then take 100 or 200 or 300 healthy young adults, give them the vaccine, and deliberately expose them to COVID. That'll tell you a month after—if you believe you may need two shots a month apart—that you wouldn't even have to do that to find out it works, just to find out the best way of doing it.
Expose them all to COVID. You have some rough idea from external evidence how likely people exposed to COVID are to get it. If almost none of them get it, you know the vaccine works. Now you can keep getting better information as you go on, but that's enough. That means we could have had some of the vaccine in maybe two months instead of 10 or 11 months. How fast they could expand production and how many would be willing to risk it, I don't know.
The impressive thing is not just that the US didn't do it. We know the FDA is crazy that way. But nobody did. I don't really know for certain what happened in China and Russia, but at least of the Western countries, none of them was willing to do challenge trials. That's a case of enough common cultural elements. One of which is: better to have 100,000 people die from not doing something than to have 10 people die from doing something. Nobody will put it that way, but that's the attitude.
That's disappointing to me. You had one country which in other ways defied the rules, namely Sweden. That was a good thing. It gives you some more information, although not as good as you'd like. They didn't do very well, but we're not sure how many other differences there are among the countries that could explain it. They did much worse than the other Scandinavian countries, but not much worse than Italy or Spain. I think better than Italy and Spain, I'm not sure. People have different explanations for these different outcomes. In a sense, I wish the world were less uniform in that sense.
If I think about aging, for example, I can easily imagine the FDA requiring all sorts of lengthy testing for medical treatments to stop aging and maybe even trying to block them. One of the effects of ending aging is that the Social Security system goes bankrupt very quickly, unless you are willing to stop paying Social Security to very old people. There are going to be other respects in which there's going to be prejudice. I could certainly imagine that some countries would see ending aging as a bad thing. I find it hard to believe that all of them would. There are obviously a lot of politically powerful people and a lot of wealthy people who are old and would very much like not to die of old age. My guess is that can't be stopped. But as I say, I would have guessed that somebody would have done a better job on COVID than they did.
Dwarkesh Patel
Robin has an interesting theory about this. He thinks that nuclear power could have been much more technically proficient by this point were it not for regulations. The way he explains how it's been globally curtailed is in a blog post titled "World Forger Elite" or something like that. I had him on my podcast if you want to listen to that part of the conversation. His idea is that the elites across the world who control regulatory agencies and policymaking talk to each other. Just like any profession, they're deeply imitative and coveting of each other's approval. Just as journalists want the approval of other journalists and intellectuals want the approval of other intellectuals, policymakers from different countries want the approval of EU regulators and FDA regulators.
This actually brings up a question I wanted to ask you about legal systems very different from ours. In your most recent book, you talk about the vast diversity of legal systems throughout history. Yet when I look out at the world today, you have some exceptions, maybe Sharia in the Middle East and so on. There is nowhere in the Middle East with the possible exception of the Islamic State, which doesn't rule anything, which actually follows Sharia.
David Friedman
The Saudis are the closest, as far as I know. But mostly countries that talk about following Sharia have a government statutory legal system, some of whose rules are borrowed from Sharia. The real traditional Islamic legal system was not legislated by the state. It was a system sort of like common law in the Anglo-American system, in which the legal rules were worked out by scholars based on religious sources. The judges were appointed by the state, but the judges didn't even have to be legal experts. They could rely on the Muftis, who were reputationally supported legal scholars, to tell them what the law was. In principle, that system had separation of state and law. States always poached on that in various ways, but that was the basic idea. No one does that now.
Dwarkesh Patel
So just to finish my question, what is the explanation for why you don't have any country that has marketable tort claims or where the plaintiff has to bear some risk for filing a suit if it doesn't work out? Why is there so little diversity in legal systems around the world today?
David Friedman
Robin's answer sounds plausible. When I wrote the book, normally when people talk about comparative law, they're talking about Japanese versus European versus US, et cetera. I was deliberately not doing that. The closest I came to our system was 18th century English criminal law, and that was because I'd written an article on it. There was a bunch of really interesting stuff there, but that's really too close to our system for what I was trying to do in that book.
You're right that the modern legal systems all look like variants on the same thing. I guess there's some pressure for consistency. If you're going to be doing business with people in another country, you'd like their legal system and your legal system to be similar enough so you can make sense of them. But I think a lot of it has got to be ideology. A lot of that is because all of these are state-made legal systems.
One of the interesting questions is if you look at non-state legal systems, which there are lots in the modern world. If you think of a university's rules for its students, a large firm's rules, and so forth, I wonder if you actually looked at those. There are two chapters in that book that I didn't write, that friends of mine wrote. One is by David Skarbek on prison gangs and one by Peter Leeson on 18th century pirates. If you look at those systems, they aren't all that similar to the state ones. If you were looking at non-state legal systems, you'd find a lot more diversity. Jewish law, rabbinic law, still exists, not as a state system—Israel doesn't really have rabbinic law—but it exists for certain areas as enforced by people believing in a religion kind of system. If you looked at those, you would find a good deal less uniformity.
Dwarkesh Patel
Even when I consider universities, if I got out the rulebook for my university, University of Texas, versus the University of Tennessee, if I redacted the actual name, I doubt I would be able to tell which one is which. Is either of those very close to either criminal or tort law in the US?
David Friedman
I don't know about those cases, but I wouldn't assume that it was. I think it is probably true that the ruling groups in all of the developed world know each other and talk to each other and not wholly unreasonably say, "That country's got this rule, seems to work for them, maybe we should do it too." There are a number of cases where there's been a deliberate campaign to do that.
The one that struck me partly because of some online discussions is age of consent to sexual intercourse. Even 50 years ago, the age of consent varied a lot from country to country. At this point, there are countries where it's different, but almost everybody is something between 16 and 18. Similarly for US states. I think Hawaii was the last one that conformed. I think it was 14 in Hawaii until maybe 20 years ago. In those cases, there's actually an effort by people who think it's bad for young people to have sex, or it happens under bad circumstances, and they push the idea, and nobody else is all that interested in it.
But the more general similarity in things you were talking about, I think are more a matter of a single global culture among law professors. I'm pretty sure that when I get messages telling me that people have been looking at something I wrote, a lot of them are people elsewhere, not in the US or not in England. Sometimes they're looking at something I wrote about law, sometimes something unrelated.
Dwarkesh Patel
I wonder to what extent this has occurred to me, but maybe it is just that we have had spectacularly successful empires in the recent past. When you think of Napoleon conquering Europe and creating a rational legal system, or the countries dominated by England, maybe they were able to just enforce a legal system that became prevalent.
David Friedman
There are really two different kinds of countries dominated by England. There are the ones that ended up with a more or less English speaking and to a significant degree English ethnic population—the US, Canada, Australia, New Zealand. And then there are countries like India or a good deal of Africa which were part of the British Empire. India was affected quite a lot. One of the two official languages in India is English. But nonetheless, it had its own culture. Although it maintained the British influence to some extent, it would not be astonishing if its legal system was quite different. I think there are significant ways in which it does not fit the Anglo-American pattern.
Dwarkesh Patel
I want to ask you about your theory of how property rights and the norms to respect them arose. Feel free to explain that theory as you answer this question: why do we not apply the same commitment strategy against the government violating our rights that we do to other individuals violating our rights?
David Friedman
Any individual who did that would be worse off as a result. That's the simple answer. What you're discussing is an article of mine called "A Positive Account of Property Rights." It was really about rights in general. I just happened to write it for a conference on property rights. The same material appears in the third edition of *Machinery of Freedom* in slightly different form.
The basic argument starts out with other species. It starts out with territorial behavior in animals, which is a well-established pattern in a fair number of different species. The territorial behavior consists of an individual somehow marking the territory he's claiming, and then somehow turning a switch in his brain so that he will fight more and more desperately against a trespasser of his own species and sex the farther into the territory he comes. The critical fact is that unless the trespasser is much stronger, a fight to the death is a loss for both sides. Therefore, if one side can somehow commit himself to "I will do whatever it takes to fight you," then the other side backs off.
This is something I ultimately got from interacting with Earl Thompson, who's no longer alive. He was a UCLA economist, a very bright, very original kind of person and a lot of fun. The idea of commitment strategies is really important.
What's the human equivalent? It's mostly not geographical boundaries. It's boundaries in right space. What are the patterns of behavior, what are the lines such that if somebody else crosses them, I am willing to bear unreasonably large costs to fight them? That then gets into the idea of Schelling points. You can't just set your boundary anywhere. You have to set your boundary at something which both sides will see as a natural division.
I talk about bank robbers who have to split the loot. Each one thinks he deserves more than half, but they don't want to keep arguing until the police show up. So they agree on 50-50. 50-50 seems like a natural division, what's called a Schelling point.
What makes civil order, the reason we aren't in a Hobbesian anarchy where everybody's fighting everybody else, is that we have reasonably consistent pictures of each other's commitment strategies. I know that if I am your neighbor and I come and say, "I've decided to dump my trash over the fence onto your property, and doing anything about it legally would be more trouble than just disposing of the trash," you aren't going to accept that. Nor are you going to accept it if I say, "Instead of dumping my trash, pay me $10 a month." In terms of a simple cost-benefit calculation, it looks like you ought to surrender, but we all have a feeling for why it would be a mistake to surrender.
In that sense, you get a pattern in which people are mostly not fighting each other because each person knows here are the lines other people have drawn. You can't just draw the line anywhere you want. I can't draw the line and say, "The deal is I will fight as hard as I can if you don't pay me $50 a month in tribute," because you won't believe me. But I will believe you if you tell me, "Try to make me pay you tribute, and I will do whatever I have to to stop you."
My definition of government is that organization against which we drop our commitment strategies. If I drop my commitments, it depends on other people seeing the difference, seeing that a government is a different kind of thing. If I drop my commitment strategy against you, other people will observe that and say, "Look, he's a wimp. We can push him around too."
This is my discussion of why the UK was willing to send their only aircraft carrier almost to Antarctica to fight Argentina over some barren islands that nobody cared about. If they didn't, it might be Gibraltar next. If it was clear that they were not committed to defend their territory even when it didn't seem worth doing, that would be a problem.
On the other hand, if you have an organization that is sufficiently powerful, and if other people don't see giving into it as the same thing as giving into an ordinary person, then you know that if you tried to maintain that commitment strategy, you would die, because the government has a whole lot more power than you do. Therefore you don't.
I really got into that part because the question I was trying to answer is what makes something a government. I call myself an anarchist. The society I describe in *Machinery of Freedom* as an anarcho-capitalist society has mechanisms for enforcing law. It's not just that you can do anything you like. Why aren't those mechanisms a government?
And that was the question I was starting with, because some people said they are, you're not really an anarchist. And my answer is because those institutions are seen by people as having the same status as private individuals. The rights enforcement agencies are not a government because if they try to extort you, you will in fact resist—not at an infinite cost, but at much larger costs than the issue seems to be worth, as with other people. Whereas with governments, you don't.
Dwarkesh Patel
Right. And in fact, I had Professor Michael Huemer on my podcast, who has written a book called *The Problem with Political Authority*. Instead of making the positive account, he makes a normative account for... sorry, he makes a normative account for why we shouldn't distinguish between a political authority and normal individuals.
David Friedman
That's right.
Dwarkesh Patel
That book, he basically runs through all of the standard philosophical arguments for why governments have a right to do things the rest of us don't have a right to do. His basic conclusion is that the only defensible one is the belief that if they don't have the right, catastrophic things will happen. And he then spends the second half of the book arguing, as I would, that it isn't true.
David Friedman
Sure.
Dwarkesh Patel
And by the way, have you heard of Charles Murray's strategy for setting up a commitment strategy to combat government interference? In his book *By the People*. If not, you'll find this very interesting.
David Friedman
Okay.
Dwarkesh Patel
So the idea is this. You have a few people who have a lot of money set up a fund, and he calls this the Madison Fund. The entire purpose is whenever OSHA or the EPA comes to somebody that normal people can sympathize with and they say, "Here are the regulations you must abide by," maybe they'll destroy your business and so on. But maybe they're obviously unreasonable. In these cases, this Madison Fund would go through every possible legal avenue, even though they know the end result is going to be that they're going to lose the case, to increase the burdens on the EPA agents and on the bureaucracies within the EPA. Increase the cost that they had to bear for enforcing this rule. And so then the government knows, "We can't go about enforcing this rule because it's just not effective."
David Friedman
It is not effective if they were committed to do it for any victim. That is to say, I presume that regulatory agencies do actually look at how hard their target is going to fight. But given that they're committed to doing it for everybody, the regulatory agency is going to say to itself, "If we give in here, we're not going to be able to do our assigned job assigned by Congress. And therefore we shouldn't." And after all, the federal government has an awful lot of money available—much more money than your group of wealthy people do.
Dwarkesh Patel
Interesting. But if you started off one regulation at a time, and in each case...
David Friedman
That's right. If you could find some way so that it only applied to 1% of what the agency wanted to do, you might be able to pull it off. That would be an interesting project.
Dwarkesh Patel
So what's the book? *For the People*, or *By the People*, rather.
David Friedman
Right.
Dwarkesh Patel
And another variant he has on this is instead of having somebody fund this charitably, you can have, for example, an association of dentists. They all pitch in $200 a month. And then the goal of this institution is if the government tries to enforce this law, which we know is not effective at preventing harm or anything, then we will chip into the legal cost of fighting the government tooth and nail to enforce this rule.
David Friedman
So in a certain sense, this is how a private law system like the Somali system did work. That is to say, in the Somali system, the traditional Somali system, you have coalitions which are formed partly by kinship and partly by explicit contract of a group of people. Part of the deal is if somebody wrongs one of us, everybody in the coalition will help him use feud—basically use or threaten violence against the person who harmed him—to get compensated. And when the compensation comes in, it gets divided among the members of the group. Typically the victim gets a larger share. It's more complicated than I'm describing, but the basic idea is that it is a group of people who are committed to defend each of them against wrongs by people who are members of other such groups. But it's even more complicated than that because it's a set of nested coalitions where if the group you're dealing with is too strong, you expand by taking in the set of other such groups that are sort of part of your tree, as it were.
Dwarkesh Patel
Right. So it's quite a neat system.
David Friedman
I have a chapter on that in the *Legal Systems* book, and it was a lot of fun.
Dwarkesh Patel
As a computer science student, I really enjoyed that chapter because it appeals to the aesthetic of recursion. When you are looking at the most common ancestor or the most recent common ancestor.
David Friedman
What I like about it, of course, is not the kinship part, although that's very important in that society, but the fact that it's also by contract, by explicit contracts, so that you can have what they call a "pile of shields," which is a coalition which is entirely by contract rather than by kinship. And apparently that was the term they used for NATO. Because it wasn't like we were kin, but we had all agreed to defend each other.
Dwarkesh Patel
Yeah, yeah. Pile of shields. Okay, so I want to ask you, does this theory imply that the poor are the least likely to have their property rights safeguarded? Because the point of a commitment strategy is the other person knows that in some sense you're protecting this property, which is not worth the cost of retribution. The flip side of that is that the poor also have less valuable stuff to steal.
David Friedman
Sure. But "the poor" is too narrow a statement. Whoever has less of the resources relevant for conflict. It might well be that a poor lawyer would be in a pretty good position because conflict in our society might well be lawsuits, and his time is available for him, whereas a somewhat better off person with no expertise in law might not be. And after all, I discussed this in the context of the Icelandic system a thousand years ago. While wealth was useful, the main thing was how many of your kin were competent warriors, how many of your friends were. So the relevant power in that system was not mainly money, it was mainly alliances.
Dwarkesh Patel
Sure. But then it's true that in such a system, how well you are protected partly depends on how much of whatever resources are used for conflict you can command. But at the same time, how much protection you need depends on how much stuff, on how strong other people's reason to want to violate your rights are.
David Friedman
Right.
Dwarkesh Patel
Where I was going to go with that was not necessarily that you have less resources for retribution, although that's part of it, but another is you just have less stuff that the aggressor could keep aggressing upon. Maybe you just have one location where your property is housed, so that once it's taken, there's no reason for the aggressor to expect you to do retribution because you have nothing that your commitment strategy is further helping you protect. It's protecting other things you accumulate.
David Friedman
And if people are too worried about that, you will then get coalitions. You will get what you had with kinship in Iceland. You will get groups of people who agree, as in Somalia, that we will, if your rights are violated, we will help you fight, and we will then get a share of the compensation.
Dwarkesh Patel
Interesting. Okay, so I want to talk now about some general questions that are relevant to this discussion. One of the ideas that's been going around in the kind of people I talk to is this essay by a venture capitalist named Sam Altman, and it's titled "Moore's Law for Everything." The idea he has, which is relatively common, is that we're going to have stupendous levels of wealth created by automation and AI. And this wealth is going to somewhat arbitrarily flow to very few agents to the detriment of everybody else because they're going to lose their jobs. And then what's necessary is some sort of wealth tax or land tax in order to redistribute this income.
David Friedman
Is it a detriment to other people, or only that other people don't get the large gain? Because my first approximation response to all of these stories is: why can't you have a subset of the population who are doing the same things they were doing before with the same technology, trading with each other, and are about as wealthy as they were before? Now, maybe you can't because maybe all the capital gets bid away from them, but they'll have whatever their own capital is. So it's hard to see how they are worse off than Hong Kong in 1950, than any society which doesn't have a lot of wealth but has productive people. So I think it's entirely plausible—and may already have happened—that technological change will result in some people being a lot better off and some people being a little better off.
If you think about it, there's evidence. I haven't really looked very carefully at the whole issue of whether wealth in the U.S. is becoming more unequal. People claim it is. It might be true, but it's complicated because you've got to look at household versus individual. If households are getting smaller, then household income would go down even if individual income didn't, and so forth. But putting that aside, supposing it's happening, one explanation is that technological progress means that the value of intelligence is increasing more than the value of strength. That's the simplest way of putting it. And therefore, even if strength is still useful, it would then be the case that the smart techies will get a lot richer than other people. And that could easily be happening. And maybe what has been happening over the last 20, 30, or 40 years—not just techies, but doctors, lawyers, people of that sort. That could certainly keep happening.
Dwarkesh Patel
But in this scenario, if it was just that the wealthier get more, they can produce more, that would in fact be good for the poor people. Because now if more can be produced because of AI, then it's cheaper for them to buy the things that they would want to buy. I'm talking about a scenario where automation makes their jobs useless. We're talking about large segments of the population.
David Friedman
I don't understand. Their jobs are useless only in the sense that there is another way of doing the same thing. But that other way is costs are relative, right? Why can't they still do the same things as before in an hour of time? So why can't they still produce the amount of stuff? You can imagine stories where they can't—stories where somehow the people who can produce enormously more are no longer available to them to work with them in producing their things. But it's not obvious that you would expect that to happen. And at the same time, the usual comparative advantage principle is that the more different the people you're trading with are, the greater the gains to both sides of trading.
So if the result is that certain individuals who can do certain things are a hundred times as productive, and certain things only twice as productive, then someone who can't do those things can do the twice-as-productive ones and trade for the other stuff and get much more than they could before. So it doesn't seem to me that it's the pattern you would expect, except if what matters to you is relative income. If what matters to you is relative income, then it's the pattern you would expect. But if it's absolute income, it isn't impossible. But again, I don't see why can't the non-techies keep running a non-techie economy, as it were, for themselves, and then have that non-techie economy trade with the techie economy on much more profitable terms than it could have traded with anybody before, because the techie economy is much better at some things than others.
That seems to me the right first guess on that.
Dwarkesh Patel
I should write a blog post because it's an interesting idea. But then, would that analysis apply? Could you say a horse, well, a horse will retain some comparative advantage over a car, but that doesn't... If we're talking 100 years ago, right? Just consider a horse versus a car. It's not as if when the car came in, by definition, the horse has some comparative advantage in terms of some benefit you can get out of it. The horse has a much more restricted set of things it can do than a human being does. And a horse doesn't own himself. Right. So to the extent that AI can do a lot more of the things that humans can do, then could it get worrying in that sense?
David Friedman
Well, I don't know. I should think more about this question. But I'm certainly familiar with the kind of scenario people talk about. I don't know what particular author you're talking about, but it does seem to me that it seems to be much more plausible that the poor people will be better off, just not nearly as much better off, because there have got to be things that the computer is not better at. People might like personal servants, and it's very hard to make a computer a real full substitute for a personal servant. They might like nannies for their kids. They might like live entertainment as something people value. It seems to me there are going to be an awful lot of things. Even if 80 or 90% of things computers can do much better, if there's 10 or 20% they can't do, even if they can do a little bit better—"better" is a relative term—you can then keep doing those things and trade them for the others.
Dwarkesh Patel
Interesting. Yeah. Okay, so I want to thank you for your time. Just two more questions. First of all, what have you learned about economics from the Society for Creative Anachronism?
David Friedman
Huh. I've gotten a more sympathetic attitude to gift cultures. I also picked that up from reading medieval Icelandic stuff. But I should have realized it earlier, even without that, because there are certain parts of our ordinary life which are gift cultures. The standard example I give is that you have some friends over for dinner and they feel as though they ought to have you over for dinner. And if they can't have you over for dinner, they should take you out to a restaurant. And it would never occur to them to say, "Well, you took them out for dinner. Why don't we pay you 20 bucks and we're even?" So that's a gift culture embedded in our ordinary system.
But certainly in the SCA, partly because I'm playing the role of a medieval person, and at least in some medieval cultures, you have the idea that nobles don't work for money. But also the feel of it is that it makes sense to do favors and not to do things for money. Now, it isn't entirely true. My wife and I have a medieval cookbook and we sell that. But generally, long before that, I put together a collection of source material on medieval cooking, which I reduction Xeroxed four pages to a page to make it really cheap, and I was selling that for cost. It somehow didn't seem appropriate in terms of my interaction with other people in a hobby to be doing it for money.
Similarly, one of my hobbies is lapidary work, cutting semi-precious gemstones. I do it for fun. I cut more stones than I have a use for. So quite a while ago, I put up an offer on Facebook to other SCA people that if somebody was making jewelry actually based on medieval jewelry, I would donate gemstones for them. Again, it just felt more nearly right to me to do that as a donation than to say I will sell them to you. I'm not quite sure why, but having participated in gift cultures, it's made it easier to have some feel for why people do that. But I still don't really understand economically. It's a very interesting question. Maybe someone else has looked at more carefully why you sometimes have gift cultures and why you don't.
Dwarkesh Patel
Yeah, I would love to read an explanation of that, just as you made an explanation for property rights. A similar thing would be very interesting, because to forego the convenience of money as an exchange for keeping track of favors—it's very interesting why that's done.
Final question. Given the sort of futurism that you've done and the different technologies you've looked at, if you were 20 today, what field would you decide to go into in a hope to either increase innovation in that space or maybe to make a profit and start a successful company?
David Friedman
Yeah, I'm not sure I'd be very good at starting up a successful company. The obvious interesting area at the moment is encryption and cryptocurrencies and so forth. But the problem is, this is the usual problem: because it's an interesting area, a whole lot of smart people go into it.
I started out getting a doctorate in theoretical physics. Part of what happened was that I was less enthusiastically interested in the subject than some of the other people. But I was smarter than the people I was interacting with, even as a graduate student at Chicago. And that was enough of an advantage. Then I was a postdoc at Columbia Physics Department, and I wasn't smarter than the people I was interacting with. They were the professors, and they were much more interested. It occurred to me that what had happened in theoretical physics was that because at the time I went into it, it was the elite field that was being subsidized, and everybody knew it was the high-status important thing—although it hasn't really produced as much since then as it was expected to—that therefore it was overpopulated with high IQ people. That was one part of the reason I eventually left that field and became an economist instead. There are certainly some very smart people in economics, but there weren't as many of that kind of people. It wasn't as dense. So it was more likely that I was smart enough to say original and interesting things in that field.
So again, the temptation—partly because computer programming is a lot of fun, I've done it, though not professionally—it would be tempting to go into computer programming. I guess the thing I would enjoy, the thing I like thinking about, and I don't know if I can pull it off or not, would be independent computer programming to do the equivalent of inventing Tetris, and maybe to do something where you can carry it far enough so somebody will pay you a whole lot of money to take over your project and then doing it really right. Making machines that work is something that human beings really enjoy. And it's a whole lot easier with programming than it is with engineering.
I have ideas for computer programs to help teach economics, three of which went with my price theory textbook I actually wrote, but I've got lots of ideas for more. I think what would be fun, assuming I could support myself doing it, or maybe do it along with some other things, would be to take clever ideas—like what I was saying before about how you ought to be running computer dating, things like that—and some mixture of writing the programs myself for simple things and getting other people involved in such projects where the project requires more than one person. That's probably the activity that most strikes me as being fun.
Now, I've also written a few novels, and it's possible that if I had started writing novels very early, I would have been good enough to support myself and do good stuff. But my guess is not. My impression is that the people who are really good at that have talents I don't have. I think I would only do that as I do it for my own fun—I write novels that some people enjoy, so it's not worth doing. Even if I compare it to my younger son, who is a would-be novelist, he just thinks about it a whole lot more carefully than I do. I do it by intuition, and he does it by some mixture of intuition and theories of how to write novels, and he's much better than I am, which is nice.
Dwarkesh Patel
Well, anyways, Dr. Friedman, this is a great pleasure. Thank you so much for your time.
David Friedman
It was fun. Thank you. I enjoyed it, too.
Dwarkesh Patel
Bye bye.