Mathematicians should dissect AI-generated proofs to gauge alignment difficulty and timelines
JacquesThibs · x · 2026-11-22
The author argues mathematicians' most valuable contribution right now is engaging with AI-generated proofs: teasing apart which techniques were used and whether anything genuinely novel appeared beyond combining existing methods. Such analysis informs alignment difficulty and AI timelines. He criticizes the prevailing habit of flippant dunks ("wasn't in the training data, huh?") over genuine curiosity about the specifics.
More from AGI Musings
- Ex-NSA's Perlroth: AI is multiplying cyber risk, exploits now beat patches by 7 days — chrisrohlf · 2026-11-22
- Training compute cost per task may be mere cents, argues 1a3orn — if models generalize from small runs — 1a3orn · 2026-11-22
- AI Is Transforming Society Because It Works as Advertised — iskander · 2026-11-22
- Szegedy predicts proofs without formal verification won't count within a year — aran_nayebi · 2026-11-22
- Quant firms 'print money' with reasoning models — and early access looks like cheating — scaling01 · 2026-11-22
- Takeoff math: separate training compute from the compute to actually explore — 1a3orn · 2026-11-22