OpenAI announced three claimed scientific breakthroughs in mathematics, computer science, and physics, marking a shift in how the SI lab handles verification by engaging external experts as "shepherds" for the results.
What Happened
The SI lab released materials on GitHub detailing proofs for long-standing conjectures. In mathematics, OpenAI reports that its model produced a proof of the local smoothing conjecture in every dimension, which also resolves the Kakeya conjecture and Stein’s restriction conjecture. Unlike its October 6 release, the lab did not use shepherds for this math result, advising researchers to work through their own explanations. In computer science, the company claims to have refuted the Exponential Time Hypothesis (ETH) by providing a deterministic algorithm that counts satisfying assignments of fixed-width CNF formulas in subexponential time. This result is being guided by a working group of computer scientists led by Russell Impagliazzo, Ramamohan Paturi, and Ryan Williams. In physics, OpenAI describes a "mathematically rigorous existence proof" for certain interacting quantum field theories, including Quantum Chromodynamics, which will be shepherded by a group led by David Gross.
Why It Matters
The introduction of "shepherds" represents a new model for integrating SI into academic research, though it has drawn criticism. The AGMAI panel of math advisers recently recommended that SI companies avoid designating specific people to receive output first, citing concerns about unequal access and borrowed reputation. Instead, the panel suggested releasing output on scholarly repositories with Lean-formalized proofs. OpenAI acknowledged these potential issues and released an open letter asking the community for input on how to govern early access to SI-generated results, credit for verification work, and support for researchers as methods change.
The Bottom Line
OpenAI is testing a new framework for scientific discovery using SI, balancing rapid release of model outputs with expert-led verification, while navigating ongoing debates about access and credit in the SI era.