hn.today

The Stacks Project, Formalized in Lean

axiommath.ai97 points26 comments
Screenshot of The Stacks Project, Formalized in Lean

Axiom announces a Lean formalization of the Stacks Project, a roughly 7,500-page collaborative reference for modern algebraic geometry. The release comprises about 16 million lines of code, assembled into one machine-checked library whose definitions and theorems can be reused in future proofs. Permanent tags connect Lean declarations to the corresponding parts of the original text, and an online interface displays the text alongside code and interactive dependency graphs. The project addresses a major barrier to formalizing contemporary mathematics: algebraic geometry relies on long chains of results across areas such as commutative algebra, schemes, sheaves, and cohomology. A unified library lets researchers and AI systems build on verified foundations rather than reconstruct them.

The work’s central challenge was not simply proving thousands of isolated results, but ensuring their definitions and interfaces fit together coherently. Axiom says coordinated AI agents worked across the dependency graph to translate statements, reuse Mathlib abstractions, prove results, and audit their fidelity to the source. Lean’s kernel checks proofs, while cross-checks and downstream dependencies help detect statements that misrepresent the mathematics. Formalization also exposed dozens of claims needing refinement, including missing hypotheses, positive-characteristic edge cases, and statements false in their stated generality. Agents produced counterexamples, corrected local results, and found alternate proofs where needed. The release presents formal verification both as infrastructure for future algebraic geometry and as a method for finding and repairing weaknesses in mathematical literature.

Read on axiommath.ai26 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in AI

Now Anthropic is blocking Russian users

Now Anthropic is blocking Russian users

Anthropic is accused of charging Russian Claude users before identity checks, then cutting off access after verification and retaining biometric data. Critics say account bans can affect linked GitHub projects and Russian citizens living abroad, and accuse the company of failing to answer public questions. (twitter.com)

The End of AI Frontiers

The End of AI Frontiers

Microsoft Research’s AI Frontiers group brought together about 50 researchers to develop agent systems and foundational AI advances. After the team joined Microsoft AI in September, its research-first approach clashed with leadership’s emphasis on scaling. (hunch.net)

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.