Axiom Math says its AI agents have formally verified the Stacks Project, a major mathematical reference. The result is about 16 million lines of computer-checked code, drawn from roughly 7,654 pages of mathematics.
Axiom Math used AI agents working in parallel to translate the mathematical library into Lean code. The agents identified tasks for themselves within a loose hierarchy, rather than receiving detailed instructions for every step.
The Stacks Project began in 2005 as an open source reference covering algebra, geometry and number theory. Lean lets mathematicians express proofs as code that a proof assistant checks line by line. Axiom also had to preserve the connections between definitions and theorems across the library.
For companies building AI agents, the broader aim is a reusable foundation of checked knowledge for writing software and eventually doing science. Formal verification already helps engineers check chip designs, but Axiom's ambition to transfer its mathematical capabilities to software remains a bet rather than a demonstrated outcome.
A large body of checked mathematics could let future AI proofs build on verified results rather than start from scratch.
The project demonstrates a way for parallel AI agents to divide up complex, interconnected technical work.
Builders should distinguish the completed mathematical formalisation from the still-unproven promise of more dependable AI-written software.
Written automatically by the Made in 10 Newsroom from the source above.
Box unveiled Box Content Retrieval, which uses Nvidia Nemotron embeddings to help AI agents search company documents. The system will roll out to Box Enterprise Advance customers in early 2027.
Isomorphic Labs is in talks to raise about $5bn at a valuation of about $50bn, Bloomberg reported. The financing has not closed, and the AI drug discovery company's work remains preclinical.
Google says AI Mode contributed to weaker-than-expected search revenue at parent Alphabet. The AI search service monetises at a lower rate than traditional search.