AI still can't do real math — and it's not just about the writing
Scattered thoughts on how doing math feels these days
A mathematician's field notes from late 2026: AI has gone from toy to indispensable in months, yet only boosts his productivity by about 50%. The bottleneck isn't raw proof power but abstraction — AI produces proofs that start simple and spiral into tangles, and it can't build the modular lemmas that make arguments comprehensible. Pushing back with human-designed abstractions often yields shorter proofs and even stronger results.
The problem is it doesn't make any goddamned sense. Really, I think it's deeper than being bad at writing. It seems like AI is bad at thinking in certain ways that manifest as bad writing.
- sdenton4
"If you just waited a year, would some other AI produce that same proof independently?"
I had a conversation about this very problem last November, with a visiting mathematician. It's essentially signaling a deflationary spiral in intellectual work: why bother doing the work now, when it will take half as much effort a year from now?
This is perhaps always a possible stance, but it's exacerbated when some proportion of your own agency is handed over to, well, agents.
- brudgers
related
Lex Fridman recent podcast on the AI crisis in math, https://youtu.be/ZAzWPE9CItg
- timbaboon
Point 19, right in the feels :(