AI proofs: 500k lines of poorly organized text nobody would bother to verify, physicists argue
skdh · x · 2026-09-13
Physicist Sabine Hossenfelder (@skdh) exchanged barbs with data scientist @kareemcarr over the quality of AI-generated math proofs:
- She sarcastically implied he's used to higher-quality papers than she is.
- His argument: if a human proof were half a million lines of poorly organized text with zero intuition for why it should be correct, nobody would invest the time to verify it — so we'd never even know it was valid, let alone celebrate it.
The exchange crystallizes a core critique of machine-generated proofs: verified correctness without human-readable intuition.
More from AGI Musings
- AI doomsday preppers content starts going viral — adamamcbride · 2026-09-13
- Altman says AGI 'went whooshing by' as OpenAI's declaration panel is never triggered — tomchapin · 2026-09-13
- 'Glad we made it this far without regulation': a brief AI-fear debate on X — rand_longevity · 2026-09-13
- "I'd bet my career on it": open models aren't insecure, closed ones are more dangerous — basedjensen · 2026-09-13
- Ex-Anthropic researcher's exit post hits 170M views as Hubinger admits >10% doom odds — adamamcbride · 2026-09-13
- Automation's blind spot: the undocumented tasks humans quietly do — chris_j_paxton · 2026-09-13