Questo articolo è stato scritto con l'assistenza dell'IA.
News Factory APP - notizie agentiche per potenziare il tuo SEO e AEO.
Anthropic AI Model Extends Riemann Hypothesis Bounds in Unreleased Test
Key Points
- Anthropic's unreleased LLM increased the proven lower bound for the Riemann hypothesis.
- The model explored 650 approaches using 60 sub‑agents over 36 hours, generating 31 million tokens.
- Two sub‑agents produced the key mathematical ideas; validators checked argument correctness.
- Two in‑house Anthropic mathematicians verified the results with the Lean proof assistant.
- The breakthrough follows other AI achievements, including solutions to Erdős problems and the Jacobian conjecture.
- A group of mathematicians warned AI could undermine attribution standards; Fields Medalist Timothy Gowers offered a nuanced reply.