Anthropic formalizes Fermat's Last Theorem in 11 days — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic said a swarm of its Claude-powered AI agents autonomously wrote a formalised proof of Fermat's Last Theorem in 11 days, confirming the correctness of Andrew Wiles' 1995 paper proof originally announced in 1993.
- The proof runs to 13 million lines in the Lean programming language and covers roughly 29,500 intermediate theorems — over five times the combined size of all previous work in the Mathlib repository, making it the largest Lean proof ever written.
- Kevin Buzzard at Imperial College London had been running a five-year project to formalise Wiles and Taylor's 100-page proof in Lean; Anthropic's result has now overtaken that effort before it finished.
- Buzzard stated the proof "leaves no assumptions other than the axioms of mathematics," adding that autoformalisation of algebra, harmonic analysis, geometry, and number theory was achieved along the way.
- Anthropic disclosed that agents "several times" lost track of the project's state and stopped collaborating effectively, and that progress only came after adopting a human-math-collaboration tool called Prove2Me to coordinate the agents and assign next tasks.
- Human experts intervened only to issue occasional "high-level instructions," with the bulk of the proof-checking and logical chain built by agents tackling smaller chunks of the theorem in parallel.
Why it matters: Buzzard himself said the achievement means automatic formalisation of the modern mathematical literature is now within reach — a capability mathematicians said was years away. By leapfrogging a dedicated five-year academic effort and producing a Lean proof larger than the entire existing Mathlib corpus in 11 days, Anthropic has demonstrated that multi-agent AI systems can execute long-horizon, verifiable mathematical reasoning at a scale previously requiring sustained human collaboration.
Ask SkimNews
