Mathematician littmath pushes back on claims AI math wins rely on stealing work in progress

littmath · x · 2026-09-11

Responding to claims that recent AI math successes stole ideas from mathematicians' work in progress, littmath argues the evidence is very weak and that such claims mostly serve to deny or downplay real capability gains — a stance he calls dangerous for the profession's survival. He concedes the weaker point that lab papers and announcements often under-attribute prior work, while noting top scientists are compensated mainly in credit and prestige. He also worries about lab race dynamics: norms of good science were seemingly sacrificed for a frankly inconsequential mathematical prize, raising the question of what principles would be traded away when more is at stake.

Related event: Mathematicians accuse OpenAI of stealing proof ideas, sparking fierce debate(6 posts)→

Original post →

More from AGI Musings

AGI Musings channel →