OpenAI's AI Cracks 719 Math Proofs

🔥 What happened OpenAI dropped a repository of 719 math manuscripts generated by an unreleased internal model. The system chewed through roughly 4,000 open research problems, averaging three hours of ChatGPT Pro thinking compute per result. 💡 Why it matters About 42% of top-line results are already formalized in Lean — a rare peek at AI-generated math near publication grade. Targets include the irrationality exponent of π and conjectures by Mahler and Kaplansky. ⚡ Our take Impressive, but hold on: without a Lean proof, it's just a claim. OpenAI should ship the remaining 58% before anyone calls this a breakthrough.