> “AlphaZero excels at learning through association,” Zhou and Riis argue, “but fails when a problem requires a form of symbolic reasoning that cannot be implicitly learned from the correlation between game states and outcomes.”

Seems relevant.

What difference do I think there is between humans and an agent? They use different heuristics, clearly. Different heuristics are valuable on different search problems. It's really that simple.

To be clear, I'm not calling either superior. I use agents every day. But I have noticed that claude, a SOTA model, makes basic logic errors. Isn't that interesting? It has access to the complete compendium of human knowledge and can code all sorts of things in seconds that require my trawling through endless documentation. But sometimes it forgets that to do dirty tracking on a pure function's output, it needs to dirty-track the function's inputs.

It's interesting that you mention AlphaGo. I was also very fascinated with it. There was recent research that the same algorithm cannot learn Nim: https://arstechnica.com/ai/2026/03/figuring-out-why-ais-get-.... Isn't that food for thought?

How good are these types of algorithms at generalization? Are they learning how to code; or are they learning how to code migrations, then learning how to code caches, then learning how to code a command line arg parser, etc?

Verifiable domains are interesting. It is unquestionably why agents have come first for coding. But if you've played with claude you may have experienced it short-circuiting failing tests, cheating tests with code that does not generalize, writing meaningless tests, and at long last if you turn it away from all of these it may say something like "honest answer - this feature is really difficult and we should consider a compromise."

Of course it's still valuable. A real app has plenty of mundane code despite our field's best efforts.

I mean no large group is a monolith so I'm sure one can find opinions either way among tradesmen. But IMO the problem is so big that it's no longer revenue maximizing for anyone, even the workers. By some measures productivity has actually been declining for construction. If that was good for workers then we should just set them to digging a second Panama Canal with spoons.

> You've got to be careful when you say what the human does, if you add to the actual result of his effort some other things that you like, the appreciation of the aesthetic... then it gets harder and harder for the computer to do it because the human beings have a tendency to try to make sure that they can do something that no machine can do. Somehow it doesn't bother them anymore, it must have bothered them in earlier times, that machines are stronger physically than they are...

- Feynman

https://www.youtube.com/watch?v=ipRvjS7q1DI

tl;dr It is very highly correlated to U-3. The paper doesn't include 2024 in the data series but the figure the article cites, 23.7%, is very near all-time best. That's pretty deceptive framing IMO.

This judgement of course depends upon the standards of the observer and where in the US you look. Before the attack on Pearl Harbor, many elites in the Empire of Japan had spent time in America and came to view America as spoiled, decadent, and too soft to fight a long war.

That comparison does not flatter Ross Ulbricht.