Anthropic says Claude worked "largely autonomously" over 11 days to formalize the proof of Fermat's Last Theorem in the Lean programming language
AI Signal Decode
Anthropic's Claude AI model has successfully formalized the proof of Fermat's Last Theorem in the Lean programming language, a task that traditionally requires years of human effort and meticulous verification. Working largely autonomously over 11 days, Claude produced a computer-checked proof by generating 13 million lines of Lean code and proving 29,500 intermediate theorems. This achievement is notable because it offers a fully verifiable proof, addressing the long-standing challenge of ensuring the correctness of complex mathematical arguments. The use of the Lean proof assistant, which requires every logical step to be explicitly defined, highlights the AI's capacity for intricate logical reasoning and detailed execution.
The market implications of this development are significant for the fields of AI research and formal verification. It demonstrates a substantial leap in AI's capability to handle highly complex, abstract tasks, moving beyond generative tasks to rigorous validation. This could accelerate the development of AI tools for scientific research, particularly in fields like mathematics and theoretical physics, where proof verification is a critical bottleneck. The success also validates investments in large language models capable of sophisticated symbolic manipulation and long-term project execution, potentially influencing future funding and development priorities in the AI industry.
Technically, this formalization represents a major advancement in AI-assisted mathematics. By leveraging the Prove2Me platform, which provides a structured framework for managing dependencies and parallel processing among AI agents, Claude was able to overcome the challenges of maintaining state and collaboration in a complex project. The proof's reliance on only the basic axioms of mathematics and its alignment with existing formalizations in Mathlib underscore its rigor. Future developments to watch include the application of these autoformalization techniques to other complex mathematical proofs, the efficiency gains in formal verification processes, and the potential for AI to not only prove but also generate novel, verifiable mathematical results.