Parisi, Zamponi use Claude AI to solve decade-old jamming puzzle
This digest was compiled by AI from multiple sources — links to the originals are below.
Nobel laureate Giorgio Parisi and Francesco Zamponi used Anthropic's Claude AI to mathematically prove why two parameters in jamming theory sum to 1, solving a problem that stumped physicists for over a decade. Their solution, published July 1 in the Journal of Statistical Mechanics, came after Claude generated a correct core idea in 40 prompts. The result resolves a discrepancy between two independent approaches to jamming, a transition from fluid to rigid disorder.
The Jamming Problem
Jamming describes the sudden transition of a fluid system into a rigid but disordered state, like billiard balls packed so tightly on a pool table that each is locked in place by neighbors. In a 2014 paper, Parisi, Zamponi and collaborators mathematically described jamming and noticed that two parameters, a and b, always summed to 1. They could not prove why, and a separate approach by Matthieu Wyart of EPFL yielded the same relation, suggesting deeper physics was missing.
AI Breakthrough
After a decade of no progress, Parisi turned to Anthropic's Claude AI. The model first reproduced the 2014 numerical results, then Parisi prompted it to prove a+b=1. Claude generated a LaTeX file with a correct core idea, though with some errors. Zamponi, reviewing the output on an airplane, immediately recognized the idea was right. After 40 prompts, the team had a verified analytical solution.
1 source
Time · lag behind first