
Ten Claude Agents Wrote a 17,895-Line Proof for a 1904 Physics Problem
Ten Claude Sonnet 5.5 agents, working under Vals AI, produced a 17,895-line Lean proof of the seven-charge Thomson problem, verified by two independent kernels.

Ten Claude Sonnet 5.5 agents, working under Vals AI, produced a 17,895-line Lean proof of the seven-charge Thomson problem, verified by two independent kernels.

OpenAI’s AI model solved a legendary Erdős problem, shaking the mathematical world. What does this mean for the future of math?

Terence Tao ported his 1999 Java applets to JavaScript in hours using AI agents, finding only one minor bug in the agent’s output.

OpenAI’s AI has shattered a classic math conjecture by finding unexpected point configurations for the planar unit distance problem, marking a new era in AI-driven mathematical discovery.