OpenAI solves 400 math puzzles – mathematicians revolt

🔥 What happened OpenAI published over 700 manuscripts covering roughly 370 math problems in October, claiming an unreleased AI model cracked them in about three hours each. The Association for Human Mathematics fired back, calling the mass release a "power demonstration," not scientific competence. 💡 Why it matters Math rewards the path, not just the answer. Dumping 700 unverified proofs shifts the burden onto humans who can't possibly check them all – and researchers are already pulling their work offline, throttling the open exchange that math depends on. ⚡ Our take This isn't research sponsorship, it's PR with inverted burden of proof: flood the zone with PDFs and let the community do QA. That's not a breakthrough, it's a data dump wearing a lab coat.