DeepMind Paper: AI Agent Autonomously Discovers New Math Theorems
A joint paper by Google DeepMind and Cambridge demonstrates that a Gemini-based AI agent can autonomously discover new mathematical rules. The agent proved exact rules for 11 previously unsolved Busy Beaver Turing machines, marking a significant shift from answering human-posed questions to open-ended scientific discovery.
Full Report
Rohan Paul (@rohanpaul_ai) · 2026-10-22 11
New Google DeepMind and Cambridge paper shows that an AI agent can discover new math itself, proving exact rules for 11 unsolved Turing machines.
AI can move from answering human-posed math questions to open-ended discovery, shown by a Gemini-based agent that found and Lean-verified new rules for 11 Busy Beaver machines with no prior documented analysis.
It tested edge cases in a simulator, drafts informal proofs, and then proves the rules in the Lean proof checker.
