factual

AI proof of Erdos problem yielded human-usable ideas

An Erdos problem was proved a few months earlier informally rather than in Lean, and the machine's human-interpretable ideas were then used by human mathematicians in a follow-up paper to prove new theorems, which is evidence that AI explanations can be useful to humans.

factualpending

Speaker

Adam Brown

Evidence Quote

“There was an Erdos problem that was proved a few months ago now, and it wasn't just an incomprehensible set of Lean.”

Source

Dwarkesh Patel Podcast - Adam Brown – A deep but accessible introduction to general relativity— Dwarkesh Patel Podcast
Created: 11/17/2026, 6:25:52 AM

Your notes

Loading notes...