Mathematician Tony Feng: current AI is not robustly superhuman yet

littmath · x · 2026-09-16

Responding to Szegedy's essay on adapting mathematics to capable AI, Tony Feng stresses the premise is hypothetical: even a Navier-Stokes solution wouldn't make current AI robustly superhuman, especially at asking questions. He suggests making PhD defenses rigorous as a first step.

Original post →

More from AGI Musings

AGI Musings channel →