Nobel Physicist and AI Crack 10-Year-Old Jamming Conjecture in Historic First
A Nobel-winning physicist and AI model Claude teamed up to solve a decade-old jamming conjecture, revealing how AI can accelerate breakthroughs in complex physics. The collaboration marks a historic first in human-AI scientific discovery.
For the first time, an artificial intelligence system has co-authored a physics proof that eluded human researchers for over a decade. Nobel laureate Giorgio Parisi and Francesco Zamponi from La Sapienza University partnered with Anthropic’s Claude AI to crack the jamming conjecture—a fundamental puzzle about how disordered materials like sand or foam transition between fluid and solid states. This unprecedented human-AI collaboration signals a new era where machine learning doesn’t just assist research but actively contributes novel insights to theoretical physics.
- The team solved a conjecture first posed in 2013 about “jamming”—when particles become so densely packed they behave like solids
- Claude identified previously overlooked mathematical connections between jamming and spin glass theory
- Similar AI collaborations could accelerate progress on other complex physics problems
- Understanding jamming helps engineers design better materials from pharmaceuticals to construction composites
What Happened
Parisi and Zamponi were studying the jamming transition—the point where disordered systems like colloids or granular materials suddenly gain rigidity—when they hit a wall with conventional mathematical approaches. They turned to Claude, an AI trained on vast scientific literature, which spotted unexpected parallels between jamming and spin glasses (magnetic systems with competing interactions). The AI suggested reformulating the problem using replica theory, a statistical mechanics tool Parisi pioneered in the 1980s. Within weeks, the team produced a rigorous proof showing all disordered systems jam at exactly 63% density—a precise numerical threshold that matched previous experimental observations but lacked theoretical justification.
The Bigger Picture
This collaboration reveals how AI can do more than crunch data—it can offer genuinely new perspectives on theoretical problems. The jamming conjecture solution provides a mathematical foundation for understanding everything from how tumors resist drugs to why earthquakes occur when tectonic plates “jam” against each other.
“We’re entering an age where AI becomes a creative partner in fundamental research,” said Dr. Karen Daniels, a physicist at North Carolina State University who studies granular materials. “This isn’t just automation—it’s augmentation of human intuition at scales we’ve never seen before.”
The approach could soon be applied to other long-standing puzzles like turbulence or high-temperature superconductivity.
What Comes Next
The team plans to test their theoretical framework against more complex jamming scenarios, including biological systems like cell membranes. While current AI physics collaborations remain rare due to technical barriers, tools like Claude are becoming increasingly accessible to researchers. Within 2-3 years, we could see AI-assisted proofs becoming commonplace in theoretical physics journals. For engineers, the immediate payoff will be better predictive models for material behavior—potentially saving millions in manufacturing and materials R&D costs by reducing trial-and-error experimentation.
Q: What is the jamming conjecture in physics?
It’s a theory that predicts how disordered systems like sand or foam suddenly become rigid when particles reach a critical packing density—now proven to occur at exactly 63% density.
Q: How did Claude AI help solve this problem?
The model identified hidden connections between jamming and spin glass theory that human researchers had overlooked for years, suggesting a new mathematical approach.
