OpenAI unleashes hundreds more math results upon a field already in shock

OpenAI published 372 results from an unreleased internal model, each said to solve or substantially advance a major open problem in math or computer science. Many come with Lean formalizations, a programming language that lets a computer verify a proof's logic, so those are nearly certain to be correct. Mathematicians say the work still needs scrutiny, since OpenAI shared no prompts and only average compute costs.

Summary written with AI by HowTechMoves, based on reporting by Scientific American.