2025-06-21
moonlapse vertigo
Dwarkesh interviews Arthur Kroeber on Chinese-US relations. There’s this idea that a lot of people have that social media is an unaligned AI. But although this sounds cool, it’s mostly untrue: social media might cause some bad outcomes, but it’s designed for a specific purpose for which it is meant to be aligned. For someone living in the west, the better analog to unaligned AI is actually the CCP: with similar but orthogonal values, a strong focus on self-preservation, and the cause of many job losses. Read abstractly, much of alignment policy research could just as easily be about how to deal with foreign powers (or aliens). I wonder if Dwarkesh himself realizes this; if so, it explains why he’s so drawn to China content from his start focusing on AI.
The extent to which you buy this lens explains why some people are concerned that “the Chinese saver is subsidizing foreign importers”. It sounds like a good deal, and it’s a misapplication to assume that China is trying to run the Uber model of charging monopolistic prices once the competition is eliminated. If they could get their factories to collaborate on prices, they wouldn’t be racing to the bottom in the first place. The danger isn’t that you will be ripped off, but that if you are entering into a patron-client relationship, the receiver of patronage will need to do what the giver demands, or risk being cut off. Fears of loss of control and gradual disempowerment apply to both China and AGI.
Anyway, one topic they spend a lot of time on is the development of the Chinese auto industry, and one suggestion given by Arthur is that China companies should be allowed to open US branches. This is both similar and opposite to my own suggestion that the CCP should let Chinese citizens invest in the US stock market. Neither is something that either side is currently open to. But I mention it because Glenn Luk has an interesting take on why having BYD US might not be the solution Arthur describes.
Money Stuff podcast with an interesting interview with Cliff Asness on active investment.
Daniel Greco has a piece on the limitations of economics. This is actually something I asked Nicholas Decker about during Manifest 2024. In general, macroeconomics as a field seems like a collection of post-hoc explanations that rarely come to consensus. Meanwhile, although microeconomics generally works, their application is only suitable within some narrow constraints under which they were designed. The question was, if we know how to make countries rich, why are most countries still poor? The response was something like “we know what to do, but these countries just don’t do it”. But surely the various structural and political problems preventing the correct reforms are also ones which fall within the realm of economics, and should therefore also be solvable if we actually understood what is going on? In general, I increasingly believe that in basically every area where competition exists, we exist close to the Pareto frontier. This is not the argument that everything is good: while it’s advances in technology that fundamentally move the frontier, in their absence, productivity increase occur when we notice that some of the criteria we are selecting on aren’t actually things that we want.
Thomas Pueyo with the background to the Israel-Iran conflict. That a numerically inferior but technologically superior country is rolling the other one over is leading to much speculation about a war between the US and China. R. W Richey on what has changed since the Korean War, while also noting that we basically can’t predict the outcome of most fights until they actually occur. Noah Smith notes that it’s functional governance, not necessarily democracy, that makes countries good at war.
Benjamin Riley with more context on the idea that Plato was against writing.
National Geographic on the Denisovan skull finding. It’s interesting because the generated picture they provide looks more southern than northern Chinese to me.
Owl Posting with highlights from comments on his endometriosis article. Also, Niko McCarty biology linkthread, and Cremieux on a potential stem-cell therapy for diabetes.
David Deek notes that population decline primarily comes from reduced family size, not childlessness. This is my hypothesis for why population decline is so hard to solve, because it implies that the most workable solution, limiting childlessness, is actually something which we’ve never had to deal with previously On that note, Stephanie Murray natalism linkthread.
Sam Enright linkthread.
