Physicist says GPT-6 Astra helped him give a candidate lattice definition of the Standard Model
OpenAI证明了「标准模型」!物理学家直言:GPT-6 Astra全程主导
In the first guest post on OpenAI's science blog, University of California, Santa Barbara physicist Nathaniel Craig describes how he worked with GPT-6 Astra to give a candidate lattice definition of the Standard Model of particle physics; Astra suggested a specific method for mapping force fields onto a finer grid and led the completion of the detailed proof. The work addresses the doubling problem for chiral fermions on the lattice and the phase alignment rules for the complete Standard Model on a grid, the latter having been unresolved for more than 25 years. Craig's idea is grid refinement, that is, continuously mapping the same set of force fields onto ever finer grids, comparing layer by layer the left-handed particle states allowed on adjacent grids, and accumulating phase corrections; Astra proved that the comparison results on finer grids converge to a definite answer. Lattice field theory experts Maarten Golterman, Andreas Kronfeld and Yigal Shamir reviewed the proof and raised questions. The related preprint has been published (arXiv 2610.09093).