OpenAI's Next-Gen Model Math Proof Rebuked by Human Mathematician in 24 Hours
JFPuget · x · 2026-08-07
OpenAI previously claimed its next-generation AI model solved 10 world-class mathematical problems, including a disproof of the Connes Rigidity Conjecture backed by 37,000 lines of Lean 4 code. However, within 24 hours, mathematician J. L. Nielsen from the University of Kansas published a paper refuting this conclusion.
Nielsen traced through the codebase object by object and pointed out that the AI's counterexample is invalid: one of the constructed groups fails to satisfy the ICC and Kazhdan property (T) conditions required by the conjecture. While the code passed formal verification, it deviated from the higher-level mathematical definitions. This incident highlights the continued necessity of human review in complex AI-generated scientific results.
More from Models
- GPT-5.6 Rewrites Triton Kernels to Cut Serving Costs by 20%, Funding Luna Price Drop — JeremyCMorgan · 2026-08-07
- Rumor: Gemini 3.5 Pro is a disaster, possibly triggering DeepMind restructuring — Neurogence · 2026-08-07
- Users Report Claude Opus Degradation: Over-Engineering and Constant Corrections — randal_olson · 2026-08-07
- Closed Models Often More Expensive Per Task Due to Token Inefficiency, Says a16z Partner — davidyin44 · 2026-08-07
- ProgramBench Eval: Gemini 3.6 Flash Sets New High in Binary Reverse Engineering — jyangballin · 2026-08-07
- MiniMax H3 Open Weights Details: 2K Path API-Only — EntireBig7258 · 2026-08-07