AI · Official
1,000 AI Agents Reproduce a Proof of Fermat's Last Theorem in a Safe Setup
Anthropic Claude blog (mirror) ·
At a glance
A team of 1,000 agents has recreated a proof of Fermat's Last Theorem. The experiment ran using only general tools, without humans directing the work. The article describes it as a small-scale test of whether multiple AIs can complete a long task together while safeguards are in place.
Key points
- The AI team took 28.7 hours and used one-third fewer output tokens (units of text generated by the AI) than a previous experiment.
- Each agent worked in an isolated virtual environment with no internet access, and a classifier checked commands and file changes before they were executed.
- The team was given only 60 statements to prove, a one-sentence summary of the classical proof approach, and 47 previously proven lemmas.
- As the number of agents grew, a single shared computer could not handle the load, so each agent used its own virtual machine.
Why it matters
The experiment shows that multiple AIs can split up a complex problem without detailed human instructions. It also shows how safety safeguards need to be designed.
Original source
Safely scaling multi-agent teams with Claude Fable 5.2
Anthropic Claude blog (mirror) · Nov 9, 09:00 KST
Read on Anthropic Claude blog (mirror) ↗AI summarized the original article. Check the original for accuracy. Copyright belongs to Anthropic Claude blog (mirror).
More from this field
- Google Brings Duplex Calling to Gemini, Letting AI Phone Businesses for You
Google has moved its Duplex phone-calling technology into the Gemini app, so Gemini can now call businesses for you. The system has changed a lot since 2018, shifting from narrow custom models to a single stack built on Gemini.
Google Research blog · Nov 21, 08:30 KST - Google Releases Project Genie, an AI That Generates Explorable Worlds
Google has released Project Genie, an AI tool that creates interactive worlds you can explore. The video shows how to access it, walks through a generated world and a T-Rex demo, and discusses why this early technology matters.
The AI Advantage · Nov 21, 08:21 KST - Anthropic Employee Says Claude Helped Prove Riemann Hypothesis
An Anthropic employee and mathematician, Levent Alpöge, says he used the company's Claude chatbot to create a detailed proof of the Riemann hypothesis over about six weeks. The proof has not yet been checked, and it is over 250,000 lines long in Lean, a computer language used to confirm maths proofs.
BBC News · Nov 21, 08:08 KST - Cole Medin Live-Streams His AI Software Factory Turning Ideas Into Working Software
Cole Medin runs his AI software factory live, taking rough requests from viewers through planning, building, and checking until working software runs on screen. He explains how it works and what he learned building it. The stream also shows a game the system built.
Cole Medin · Nov 21, 08:08 KST