Weekend Reading & Selected Links

Happy weekend! Here are some links to things I've been reading or watching that you might also enjoy:

1. My new podcast episode, with Bryan Caplan. Four excerpts from the conversation below.
2. New talk on learning and AI, by friend of the pod Andy Matuschak.
3. 'Watch Auckland transform from zoning reform'. Before and after photos.
4. The AI reading list Ilya Sutskever gave to John Carmack. Via Keshav Chan.
5. Ben Todd's Google Doc on AI for epistemics.
6. The physical dimensions of intellectual projects.

Have a great weekend,‌
‌‌
‌‌
Joe

Excerpts from my podcast with Bryan Caplan

1. On density and fertility

WALKER: Okay, let's talk about fertility. So in a way, I would love for it to be true that housing unaffordability was an important cause of declining fertility rates, as you suggest in Build, Baby, Build, because that would offer up a more tractable solution than some kind of harder to pin down cultural explanation. But there's a lot of research showing that there's an association between density and reduced fertility. And a couple of very clear examples of this would be Tokyo and Seoul. So what do you make of this result? Doesn't population density lower fertility?

CAPLAN: So here's what I'd say is that these papers are making a good point. It's one that we should be thinking about, but they are really jumping the gun because normally high density and high prices go hand in hand. We'd really need to be finding places where there's high density but housing is really cheap – and that's really hard to find. Why? Well, normally, anytime you actually get a city, then regulation is going to be at least fairly strict. And so we just don't get to actually observe this alternate universe.

WALKER: Is there anything that fits the bill there? Are there any examples of cities that are both high density and have low prices?

CAPLAN: So Japan has had flat prices, but they're still actually high, especially Tokyo. They're still high. So, yeah, I don't know of any good exceptions. There may be some somewhere on Earth. But that would be where you'd want to start. Definitely what we do know is there's basically places of similar density but different prices. So that is what you need for minimal social science. And as far as I can tell, the people that are getting worried about this just haven't done that basic homework of trying to disentangle price from density because those are logically different things. You can imagine an area that's very dense and yet the housing is cheap. So that someone that wants to go and have a lot of living space can get it. And we really just don't see that. 

2. How likely is YIMBYism to become left-coded?

WALKER: Now that you've raised politics, I'll ask my political questions. So Operation Warp Speed was successfully run by the Trump administration. Trump boasts about “his vaccine”. But then somehow vaccines become left-coded, and so it becomes a Republican thing to not take a vaccine. Right now, housing abundance and YIMBYism seems more or less bipartisan, if you ignore the far left and the far right. How likely is it that YIMBY becomes either left-coded or right-coded? And which of those options is more likely?

CAPLAN: Yeah, good question. So I say right now, the two actual centres of YIMBY: there's largely left wing activists in major blue cities, and then there are just red states that are YIMBY by default where they haven't really turned it into a philosophy – it's just always been that way, and they don't have the same kind of resentment against construction. 

WALKER: Like a Texas.

CAPLAN: Yeah, like a Texas. So, basically, the places that actually have low regulation are clearly red states. The places that have an activist movement who want to have less regulation are basically blue cities and generally blue cities and blue states. 

In terms of what this means for the future, in the book I just express my dream, my hope, that it could be bipartisan. But, yeah, I think you are right that if something becomes popular, it will tend to go to one side. 

I would say that, strangely, the most likely outcome is that because the actual vocal activists tend to be left wing, that is enough to go and give it a left wing coding, because they're the ones that'll be on TV talking about it...

The low regulation is in the red states, but the falling regulation is in the blue states. You get more media attention for the change than for the level, which is another reason I think that it will be thought of as one of those weird things they're doing in left wing states. Even though the truth is that most NIMBYs are also left wing. And these places – definitely towns with the very strictest regulation – are generally very left wing towns and places, places like California.

WALKER: Yeah, yeah. So at the moment, Ezra Klein and Derek Thompson are writing a book on supply-side progressivism. And I don't know, but presumably there'll be a chapter on housing, it’ll be one of the main chapters. What would your advice to them be? Tread carefully? Maybe just make it a section rather than a whole chapter?

CAPLAN: No, I mean, honestly, I almost always tell people to max out. I just figure that the people on your side read you to such a greater extent that you should never worry about improving your own side. Always try to make your side as good as it can be, and especially combine it with saying there's something we can learn from the other side. That last add-on does tend to make it hard to really persuade your side. But that's why it's great that someone as high status as Ezra Klein is going to be the voice, because it's like, “Well, do we go and kick him out of the movement or do we listen to him?”

3. Putnam's famous 'E pluribus unum' paper overstates its own result

CAPLAN: There's a famous paper by Robert Putnam. I think it's the only paper that he ever did on trust where he went and did this actual quantitative work. This is one where I did not find an arithmetic error. Rather, I found that he just had a table where the arithmetic implied that the rest of the paper was greatly overstated. Because it wasn't like he went and did it, and then he said, “Okay. And then this shows there's this huge effect on trust.” He just put up a table, and then he just said, “And so we see that this is really horrible.” And it's like, well, is it? I just went and said, alright, so according to his paper, what is the effect of moving from our current level diversity to maximum level? And it was just a microscopic change. 

WALKER: Oh, that's so interesting. I know the paper. But I’d just heard the punchline.

CAPLAN: Right. So you can link to it in the show notes. So there's a lot of verbiage about how, “Oh, my God, it really pains me to accept this harsh truth that diversity is really bad for trust.” Then you look at the actual table and it's saying that trust is falling by a microscopic amount in the most extreme scenario. So then it's like, it's weird that you would do that. You would think that if you were really disturbed, you would go and multiply the numbers out… And it's like, well, maybe you weren't really disturbed as you said, or maybe you just got so worked up over the big picture that you didn't go and double check that the numbers actually were in line with your adjectives.

4. Japanese urbanism

WALKER: A couple of questions on urbanism and architecture. So you recently took a trip to Japan, right?

CAPLAN: Yeah.

WALKER: Was there anything on that trip that you learned or that surprised you about Japanese urbanism?

CAPLAN: So, one thing you can definitely see is they just barely worry about historic preservation because their cities just got torched during World War II. They just don't have much that's very historic. So, yeah, if you know the horrors of the war, Tokyo's basically burned to the ground. 

WALKER: The firebombing.

CAPLAN: So they had a long tradition of building in wood, and then also horrible firebombing and everything else. So they just don't have the same issues with, “We’ve got to preserve the past.” I mean, a lot of people also say there's this cultural thing where, because of Buddhism or Shinto, they just have the idea of reviving things. For historic buildings, they will actually periodically go and take them down and then rebuild them from scratch because it's made out of wood and the wood won't last forever.

There's this very famous shrine in Nara where it used to be – it’s still – the world's largest wooden building, but I think my book said it used to be fifty percent bigger. And they said, “We're just not going to go and build it the same size.” So that was something where it's like, wow, even for a tourist site, you'll go and do something different based upon, like, “We didn't have the money,” or… I don't know what the rationale was. So that's one thing. 

You can definitely see the benefits of mixed use there, because life is just really convenient in Japan as a result of having almost no regulation of “Can I put a store in the ground level of the building?” I'm going to admit I'm not an expert in Japanese regulatory law. I do know that it's at least a lot easier for a residential home to operate as a small business in Japan and just to go and hang out a little shingle and say, “We live here, and also you can get some sushi and just knock on our door.”...

In terms of density, you can see that all of the heralded wonders of mass transit, which I'll say even in Europe, are really overblown and it's not that good. Ed Glaeser has a nice paper showing how the main reason why people like cars is that cars are faster almost everywhere. If you have any doubt, you can just go and start typing in “transportation times for car versus train”, even in Europe, and you'll see normally, at best, it's a tie. With very few exceptions, it's a tie. And for anything non-standard, car is much faster. Even in Europe, even in Germany or in the Netherlands or Belgium or Luxembourg, where people are so enthusiastic about how wonderful the mass transit is. 

And the Glaeser story is really pretty similar, namely, look, with mass transit, there's always the fixed cost of walking there, waiting and then you get off and then you have the final thing, or maybe there's transfers. With a car, you can just go out your door and drive there. And by getting rid of those fixed costs, you are able to go and save time for almost every kind of journey, unless there's horrible congestion. When people are saying, “The US government is really encouraging cars.” Is it though? If they did peak load pricing for traffic and for parking, then cars would be a lot better. I'd be happy to drive right into New York City if it cost me $40 but I knew that I could just go and pay market price in order to have no congestion and be able to park.

So I’d say the actual package of encouraging cars in the US, once you realise that not having tolls is anti-car… People think tolls are anti-car. No, no, it's not having tolls that is anti-car, because if you had tolls, then you could use the technology. 

But anyway, so Japan is a place where you can go and put in the times for driving versus mass transit. And you say, okay, the mass transit is actually absolutely faster in Japan, because it's not just the density, it's also just their incredible efficiency and conscientiousness of the Japanese. Their cultural differences are quite shocking, when you're there and you just see, wow, this would not work in other countries. In other countries, they would just be screwing up a bunch of things. And here in Japan, their trains really do work.