AI's Rapid Math Progress Has Researchers Losing Bets
ChrSzegedy · x · 2026-07-16
Former prominent Google researcher Christian Szegedy stated that, technically, he might be close to losing a bet with François Chollet. They wagered on whether AI could solve more than 10 unresolved mathematical conjectures (that had been open for a decade) and automatically formalize 50 textbooks before June 2026.
According to the cited tweet, users have already leveraged the GPT series models (primarily 5.5 Pro, as well as 5.2, 5.4, 5.6, etc.) to successfully propose solutions to 19 Erdős problems and complete multiple partial proofs. This marks a substantial breakthrough for AI in the field of advanced mathematical reasoning.
More from AGI Musings
- ControlAI CEO says an international ban on superintelligence is needed to avert extinction risk — zetalyrae · 2026-07-22
- Gary Marcus says LLMs still cannot really do math on their own — GaryMarcus · 2026-07-22
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- AI may make digital work infinitely leveraged while offline life gets more human — illscience · 2026-07-22
- Better AI math could save researchers time by killing false conjectures earlier — prateekj · 2026-07-22
- AI’s economic forecasts are split by nearly a quadrillion dollars by 2035 — bittingthembits · 2026-07-22