OpenAI published a writeup plus a Lean formalization that it says proves Khot’s Unique Games Conjecture, and leading complexity theorists (including Dana Moshkovitz) are publicly studying and explicating that purported proof at major institutions. The claim combines a foundational theoretical‑CS milestone (implications for hardness of approximation) with an AI lab producing machine‑checked formal mathematics.
— If true, the proof rewrites many optimal inapproximability results and signals that AI labs can produce machine‑formalized breakthroughs, raising questions about verification, credit, and the institutional role of AI in mathematical discovery.
Scott
2026.10.30
100% relevant
OpenAI’s released writeup and Lean formalization of a Unique Games proof (per the article) and Dana Moshkovitz’s scheduled technical talks at the Simons Institute and UT Austin explaining the construction.
← Back to all ideas