Scott Alexander (Astral Codex Ten)

End quote.

43.

Just as there are stock indexes like Nasdaq or Shanghai Composite to easily track questions like how is tech doing and how is China doing, Metaculous is experimenting with prediction market indices.

I'm skeptical of their flagship example, how ready are we for AGI, which seems to be a weird mismatch of questions about how good AI capabilities are, how well technical alignment is going, and stuff like UBI.

Split between recommending better curation versus worse curation, for example something more like Nasdaq that includes so many thousands of stocks that it can't help but track underlying trends.

My list of links to publish today includes something like a dozen about DeepSeek, which now seems so thoroughly yesterday's news that I'm tempted to throw them all out.

But in case you still have questions about it, I felt most enlightened by takes from Dean Ball, Helen Toner, and Miles Brundage.

The story seems to be that DeepSeek genuinely did a great job, made extensive algorithmic progress, and was able to create an excellent AI on chips scrounged up from before the export controls hit, and mediocre chips that got through the export controls.

Along with these real reasons to be impressed, there's also a little bit of illusion at work.

OpenAI delayed announcing O1 for a long time.

Remember the rumors about Q-Star and Strawberry?

and DeepSeek was very fast to announce R1, which made DeepSeek seem closer behind OpenAI than they really were.

Most of the smart people I read say that the absolute worst response to this, from an arms race point of view, would be to give up on export controls.

If a rival has geniuses who can use resources ultra-effectively, you don't want to also give them more resources.

45.

Nils Wendel on looking for your first job in psychiatry.

Lots of good advice, but don't be intimidated.

I think I did about 25% as much work as he did, and it turned out fine.

I would have done closer to Nils' level of work if I'd been going into hospital rather than private or outpatient psychiatry.

You're more of a cog in a machine there, and you want to make sure that the machine is a good one.