Even if humans lose frontier math, physics can ground whether AI's progress is useful, argues Darpinian

Darpinian · x · 2026-09-09

Responding to Fchollet's concerns about AI frontier math, Darpinian argues that even if mathematicians mourn the death of human-driven frontier math, other math-reliant fields would benefit more than before — they can request the advances they need without needing frontier math expertise themselves.

He adds that even in a world where humans no longer understand the frontier of math, physics and other fields concerned with real-world behavior can provide the grounding that decides whether AI's "progress" is useful in direction. A pragmatic counterpoint in the debate over whether AI-generated math counts as real progress.

Related event: Terence Tao's long response to AI's Navier-Stokes claim earns Chollet's praise(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →