News
2026
-
October 2026
Interview in New Scientist: Will physics be the next field to get turned upside down by AI?
-
September 2026
Four papers accepted at NeurIPS 2026.
-
September 2026
New paper: How good are frontier models at physics? Expert re-grading shows that leading physics benchmarks are broken, and that, contrary to what their published scores suggest, frontier models are close to saturating them.
Post on X
-
September 2026
New paper on world models that learn physics: Semigroup-JEPA: Latent dynamics consistency for zero-shot physics generalization.
-
July 2026
We organized AI4Physics, the first ICML workshop on AI for physics, at ICML 2026 in Seoul.
-
May 2026
New paper, a spectral lens on optimization: Understanding optimization from activation and gradient spectra: A modded-NanoGPT study.
Post on X
2025
-
September 2025
Co-organized the AI + Physics Workshop at the CUNY Graduate Center.
-
June 2025
Awarded the 2025 Nevill Mott Medal and Prize by the International Conference on Strongly Correlated Electron Systems (SCES 2025).
-
April 2025
Massimo Capone wrote a commentary on our bipolaronic high-temperature superconductivity paper for the Journal Club for Condensed Matter Physics: How can bipolarons manage to stay slim?
-
April 2025
Gave the Yale Physics Club colloquium on the 2024 Nobel Prize in Physics: The rise of neural learning.