DigestAI news desk
hot OpenAI board member warns company is not on track to prevent catastrophic AI loss of control new ChatGPT prompts enable regional 1980s-style photo transformations in South India hot OpenAI launches Agents API beta for long-running cloud agents new Guardian: US-China AI talks needed as Anthropic reports bio-weapon misuse attempts hot OpenAI Unveils GPT‑6 Astra: Record‑Breaking 3D Rendering, Loop‑Transformer Architecture new Nscale appoints Fidji Simo to board ahead of potential $3.5B IPO hot OpenAI solves Navier-Stokes problem, sparking academic controversy over data use new Chinese AI Models Threaten Anthropic & OpenAI IPO Valuations hot gPTY: Godot‑Rust Terminal Multiplexer Enables AI‑Driven Workspace Control new IFA 2026 Highlights: High-End AI Laptops vs Budget Machines hot OpenAI Introduces ChatGPT for Financial Services new Owkin Licenses AI to Servier for Oncology Research hot Anthropic releases 150-page report on global Claude misuse and distillation new Vinyals Says AI Self‑Improvement Is Inevitable but Not an Explosion hot RTK Token Savings Debunked: Cost Benchmarks Disagree new Salesforce Launches Job-Ready AI Agents Across Sales, Service, Commerce
Research updated

Anthropic's Claude formalizes Fermat's Last Theorem proof in 11 days

Anthropic announced that its AI model, Claude, has successfully formalized the proof of Fermat’s Last Theorem into computer-verified code. This achievement marks a significant milestone in mathematical AI, as the model translated the complex 1994 proof by Andrew Wiles and Richard Taylor into 13 million lines of Lean code. The process took only 11 days, a task estimated to require approximately…

1 source primary source HN 6

Key points

  • Claude formalized Fermat’s Last Theorem into 13 million lines of verified Lean code in 11 days.
  • The task was estimated to take humans 10 years, highlighting a massive efficiency gain for AI.
  • Experts view this as a major step toward AI verifying the entire library of mathematical knowledge.

Mathematicians, including Alex Kontorovich from Rutgers University and Kevin Buzzard from Imperial College London, described the result as a breakthrough that demonstrates the rapidly improving capability of AI to handle high-level mathematical reasoning. Buzzard noted that this task was an order of magnitude more difficult than previous AI formalization successes, such as the certification of Maryna Viazovska’s sphere-packing work earlier this year.

The development suggests that AI will play a crucial role in verifying existing mathematical knowledge and potentially discovering errors in established results. Experts believe that at the current pace of progress, AI systems may soon be able to scrutinize the entire library of mathematical proofs, fundamentally changing how mathematical validity is established and checked.

Read the original at Nature Machine Learning primary source Open source ↗

Coverage and discussion

1 source
Topics · follow one to build your own front page
AnthropicClaudeAlex KontorovichKevin BuzzardDaniel LittAndrew WilesRichard TaylorMaryna Viazovska

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us.

Comments

via GitHub Discussions

Related stories