41 Comments
User's avatar
Fallingknife's avatar

Dying or mean reverting? If we had the same number of social science and humanities phds as we did in 1960 we would have about 1000 of each, and I see absolutely no reason why we need more per capita now than we did in 1960.

cp6's avatar
Oct 27Edited

I agree. “Dying” means something is going down to zero and, once there, will stay at zero. That is almost certainly not happening to the humanities.

The proper word here is “declining”. The humanities are on a downswing right now but people still want them. We still need people to teach history and literature, make good art, lead us and inspire us. Those needs are not going away. As long as there are humans there will be humanities.

Tim's avatar

Dying. The chart is not clearly labeled but it's showing the number of new dissertations defended in each year. If you click through to the study, the authors examined over 443,000 dissertations over the study period.

Falous's avatar

I had a very similar thought.

The Ghost of Tariq Aziz's avatar

The collapse in new humanities PhDs is a good thing. It is unfair to students to admit them to programs where their chance of getting a tenure track position is one in a thousand. They end up spending the prime years of their lives pulling out their hair in order to become less employable. I question the judgment of the tenured professors who made this decision way back in the nineties. It seems profoundly selfish to me. The class size of the top twenty programs should be commensurate to the actual supply of jobs.

cp6's avatar

It’s not just social media - it’s also games and short-form video apps on smartphones.

All three of these are addictionware, designed to suck people in and keep them hooked. I don’t know how many smartphone games you’ve played but most of them are amazingly bad, designed to addict you and make you feel compelled to spend a certain time every day playing.

It was hard to kick my addictions to those games, but I managed to do it and I’ve stayed away since. I learned a lesson: be the boss of your phone, don’t let it be the boss of you!

David Khoo's avatar

2. The graph you have for investment is only for the "construction" component of investment, the buildings and powerlines, which are only a small part of AI investment since it won't include the GPUs, racks, gas turbines, etc. This should be obvious because the orders of magnitude are off. US AI investment in 2026 isn't 50 billion; it's more like 500 billion. Alex Tabarrok's post quotes another graph from Krugman with his estimate of all tech vs non-tech investment, which looks much more convincing.

But ultimately any story about "crowding out" relies on what counterfactual you select. This is the gist of the disagreement between Tabarrok and Krugman, and your timing argument against. Instead of asking if there's crowding out or not, instead just ask if in another world where there weren't 500-1000 billion of AI investment in one year alone, whether interest rates would be higher or lower. That's the key question. I think the answer is non-trivial, but I tend toward higher.

Stephen C. Brown's avatar

What about renewable energy investment? That’s a similar industry likely to have soaring future demand, not like the unglamorous infrastructure investments. I wonder if lowering industry capacity usage could map onto these investment curves?

Don Geddis's avatar

Keep fighting the fight on rent control! It feels enormously intuitive to the vast majority of voters. The political debates are all about whether society should "care" about poor people, or else reward those who are already rich (and the poor get what they "deserve"), in some kind of zero-sum game.

The fact that economists know that rent control is bad (on average) EVEN FOR the very populations it is ostensibly designed to help ... isn't even in the political conversation. We're such a long, long way away from voters having an informed, educated opinion on this subject.

Rie Yano - Coral Capital's avatar

We are honored and pumped to work together Noah!

Worley's avatar

I read this and laughed -- the phrase "honored and pumped" is such a perfect combination of ultra-Japanese formality and ultra-American informality! Clearly Coral Capital understands messaging well.

Joseph's avatar

Great roundup, especially the Japan housing story

nic p's avatar

I've been working on an electronics project with a friend for the past few months and it is shockingly difficult to do non-china sourcing. I'm hoping that as China's low-tier manufacturing declines we'll start to get some of that component/pcb manufacturing in Vietnam/Malaysia etc

CW Morris's avatar

Google "comparative employment prospects of philosophy majors". Many fields in the humanities deserve their bad reputations, but not philosophy.

Kazimir Czerwinski's avatar

Congratulations on your new role! It sounds exciting and fun, all the best to you (and your rabbits!)

Kazimir

Ted's avatar

Noah, I love the Bebop shirt!

See you space cowboy

Hoff's avatar

2. The bigger factor in the crowding out is that data center demand has driven up the price of engineering & construction labor, or locked that labor up completely in long term contracts.

This isn’t just a function of frontline construction workers, plumbers, electricians, etc. - construction management is also constrained. EPC firm capacity is significantly more expensive or cannot be had at any price.

These prices haven’t corrected because the underlying labor pool is aging and structurally limited, and retraining takes time. Individuals may also be reluctant to retrain if they suspect the AI spending boom is transient.

The timing for this causal explanation lines up much better than the interest rate explanation.

Worley's avatar

I suspect that part of the cause is Trump's immigration crackdown. Construction notoriously employs a lot of Latinos and I would expect that up until recently, a lot of unauthorized immigrants were employed in construction (creating the future of the United States). Anything that decreases the construction labor pool is going to reduce construction investment.

Charlie Hammerslough's avatar

I recall that the building of Narita Airport was met with violent opposition from the communities whose land was expropriated.

Sassy's avatar

It was, which really put a chilling effect on infrastructure. It's the reason why the Shinkansen has a slow bottleneck between Omiya and Tokyo. The first connection was built with a slower top speed as appeasement, and the second connection was never built, resulting in permanently less competitive high speed rail to Tohoku/Hokkaido and the Sea of Japan side.

Peter Lerner's avatar

Noah, Is The Weeb Economy available in English? I can't find it anywhere to purchase.

Rochelle Kopp's avatar

Using bureaucrats instead of courts to do environmental review is not necessarily the great idea you think it is. Because local bureaucrats don't necessarily have much incentive to listen to concerns of the community, especially when they are in cahoots with the developers (as often happens in the case of large redevelopment projects particularly ones using parkland). And faster often means prioritizing developer timelines above community input. If you are interested in exploring this topic further, I would be happy to introduce you to Professor Harashina, pre-eminent scholar of environmental assessment in Japan and former head of the International Association for Impact Assessment (IAIA). He can give you an earful on all the problems he sees with how environmental assessments are conducted currently in Japan. I can also share my observations related to development projects at Jingu Gaien and other parks in Japan, where environmental assessments are mere check the box formalities, or are not even conducted at all.

I will also note in the Cremieux article you shared, where he shows two aerial before and after views, almost all the trees disappeared between the first and second photo. Don't tell me that some great environmental review was happening there...

Fallingknife's avatar

Better the corruption of developers bribing officials that actually gets things built than the corrupt process that shovels money into the pockets of useless lawyers, bureaucrats, and unions and prevents things from getting built.

Rochelle Kopp's avatar

Why does it have to be either or? Both could be improved.

Sassy's avatar

> almost all the trees disappeared between the first and second photo

Yet Japan is one of the most forested countries in the world. And even my car free access to natural recreation activities from the suburbs of Tokyo is orders of magnitude better than it was in the suburbs of California or Michigan.

> problems he sees with how environmental assessments are conducted currently in Japan

The biggest problems are still in delaying or killing beneficial projects like solar and wind farms.

Rochelle Kopp's avatar

Japan may be very forested but it is losing trees in urban areas rapidly, from a small base, which is exacerbating the heat island issue. A good article on this recently: https://asia.nikkei.com/spotlight/datawatch/throwing-shade-japan-s-urban-tree-coverage-dwindles-bucking-global-trend If you live in the suburbs of Tokyo you are likely in one of the places that is experiencing significant canopy cover decrease -- for example 40% in Suginami-ku over ten years. Nice that you can get to green recreation areas outside the city without a car but greenery in the urban environment itself is also important.

As for solar power it isn't that they aren't getting built. There are subsidies to encourage them and there are so many now that once they are built the power from them isn't being utilized. https://www.linkedin.com/pulse/when-japans-solar-power-has-nowhere-go-james-jim-hoadley-g0kce/ Here is a good article on environmental issues with renewables in Japan https://jamesryland.substack.com/p/trees-in-japan-being-cut-down-again?utm_source=post-email-title&publication_id=2575715&post_id=221782034&utm_campaign=email-post-title&isFreemail=true&r=brb2&triedRedirect=true&utm_medium=email (I agree with him that more offshore wind would be a good idea and also think Japan could do much more rooftop solar rather than the giant projects). Also there are significant documented problems with the large solar projects that cut down many trees, including increased risk for landslide, which is an argument for more environmental assessment not less I would say.

Sassy's avatar

> which is exacerbating the heat island issue

The Nikkei Asia article is written by someone clearly out of touch with reality. It supports fucking road widening for fucks sake. The BIGGER ROADS MORE TREES nonsense that leads to US suburban hell.

Citing data for Calgary shows the writer is clueless and haven't actually tried walking through one of the many large parks in Tokyo to see if it actually feels cooler than a shady alleyway. Humidity is so high that evaporative cooling tends to do nothing to heat index or increase it. If you actually cared about the urban heat island effect, you'd just cut down all the trees to replace them with shade structures.

I do like trees though. It's nice decoration, and pretty colors that follow the seasons. Trees help especially on wider roads where they break up the vast expanse into something more human scale. Except it's incredibly difficult to make a wider road nicer than a narrower road even with tons of trees. It's hard to take anyone advocating for the massive roads cutting across western cities seriously.

It's so much nicer to walk around Tokyo than New York or London. An article that advocates for making Tokyo more like New York needs to acknowledge that and provide an adequate counterargument. And the writer expects you to take BIGGER ROADS MORE TREES without justification.

> Nice that you can get to green recreation areas outside the city without a car but greenery in the urban environment itself is also important.

It is, but the least hospitable roads I've walked along were very green from a satellite photo.

> Also there are significant documented problems with the large solar projects that cut down many trees, including increased risk for landslide, which is an argument for more environmental assessment not less I would say.

It is a risk, though I've mostly heard it from people who spew nonsense about urban planning. It's hard to take anyone who links an article in favor of mass road widening in Tokyo seriously.

Rochelle Kopp's avatar

I believe the Calgary data was shared because that is where a study with useful data about the effect of trees on temperature reduction was generated.

I don't think that citing an academic study is "clueless." You might also want to google "value of urban trees" for more information on why urban trees are important beyond being pretty.

I didn't see this article as advocating for mass road widening. The overall thrust of the article was that Japan is losing tree canopy cover and that urban planning needs to take into account shade from trees. There is mention about there not being space for trees on many roads in Tokyo, and mention that few redevelopment projects widen roads to make space for trees. So that seems like a comment limited to areas subject to redevelopment projects, and also seems to be advocating for widening the area on the side of the road where the trees are planted rather than widening the roadway itself. I don't think that means the article is advocating mass road widening.

And actually, road widening (of the roadway itself) often leads to cutting down trees in Tokyo.

Buzen's avatar

Why would a local community have more knowledge about environmental impacts than local or other bureaucrats? Most local environmental opposition is just made up impacts about traffic, parking or shadows and esthetics as an excuse to maintain property values and neighborhood “character”, not actual environmental harms.

Rochelle Kopp's avatar

I think you're setting up a false dichotomy between supposedly knowledgeable and impartial bureaucrats and self-interested NIMBY residents. The reality in Japan is considerably more complicated. Local governments are often actively promoting redevelopment projects in cooperation with private developers, while also overseeing the environmental assessment process (or deciding whether an assessment would even be conducted, often it’s not despite obvious environmental impacts). That's an obvious potential conflict of interest.

And local residents often have knowledge about their environment that bureaucrats and consultants don't, not to mention that they have to live there after the development is completed. That's why public participation is an important part of environmental assessment. Technical expertise and community input should complement each other.

Mature trees, biodiversity, urban heat mitigation, and access to public green space are real environmental issues, not just excuses to protect property values. And the usual arguments from the U.S. context about NIMBYs blocking housing don't necessarily apply to projects that replace public parkland with commercial developments, which is what I have been involved with activism about in Japan.

Japan may have some useful lessons for the US about making it easier to build housing. But that doesn't mean Japan has figured out how to do environmentally responsible development and redevelopment. The real question is whether the process is sufficiently independent, transparent, and accountable. There are significant issues with that in Japan.

Worley's avatar

> local bureaucrats don't necessarily have much incentive to listen to concerns of the community

True, but once you scrape the paint off, that's also the benefit -- the concerns of the community tend to be dominated by a conservative itch to prevent anything from being changed.

I'm reminded of an article in today's Boston Globe -- an inner suburb of Boston, Cambridge, is trying to loosen the zoning controls on building new housing. They're discovering that younger people are much more open to building ore housing and older people are generally against it. The *point* of abundance is not listening to the latter people.

Rochelle Kopp's avatar

I think you're taking the US debate about NIMBYs blocking housing and assuming it applies to every situation where citizens oppose development. I'm trying to point out that the situation in Japan is quite different. Yes, Japan builds housing, but it's also losing urban tree canopy, and public parks and green spaces are being replaced with concrete high-rises and parking lots. These are real environmental losses, not just people resisting change because they don't like it. I find it frustrating that concerns about the destruction of public green space are so readily dismissed as NIMBYism. Surely it's possible to support building more housing while also recognizing that not every development project is a good one, and that citizens sometimes have very legitimate reasons to object -legitimate reasons that government shouldn’t just be ignoring.

Worley's avatar

How do you build more housing within an existing metro area without reducing the amount of green space? ... I suppose you could replace existing housing with taller housing.

Worley's avatar

Regarding the cost reduction in AI, two points:

(1) Typically the reduction in unit cost of a product is plotted against the unit production of the product, not money invested. See https://en.wikipedia.org/wiki/Experience_curve_effect. I haven't been able to find any graph like that.

But I remember from a few decades ago an analysis that showed that the dramatically fast drop in the price of transistors was only a fairly normal drop in price *per doubling of unit sales* because the unit sales of transistors were very price-sensitive as transistors displaced other ways of controlling things. So there was a powerful positive feedback loop at work.

(2) "For decades, it looked like financial capitalism was broken — no one was thinking for the long term, everyone just cared about quarterly numbers and shareholder buybacks" isn't quite right. All of these phenomena were real, but hardly surprising in a market economy. What was really unusual was "large tech companies hoarding cash" -- the typical dynamics of large companies is that the shareholders favor that idle cash be distributed to them via dividends or share buybacks, and if the corporation doesn't, their stock price sags. But observers noted that at the time, shareholders were rewarding tech companies for retaining cash, as if they were scenting that some big change was coming that would require the companies to spend a lot of money. And it's that cash that they spent on the first wave of AI.