← Sunday, November 15 deck

02Research2 sources posting

OpenAI Opens a Science Blog as Physicist Nathaniel Craig Says GPT-6 Astra Helped Build a Candidate Lattice Definition of the Standard Model

Physicist Nathaniel Craig shares how he worked with GPT-6 Astra to develop a candidate lattice definition of the Standard Model. It’s the first post on our new science blog:

Context

Why it matters

If the construction holds up under review, it would be one of the clearest cases of an AI model doing core proof work on a decades old problem in fundamental physics, not just helping with calculations.

OpenAI posted the guest essay on November 15 as the first entry on a new science blog. Its author, Nathaniel Craig, leads the particle theory group at the University of California, Santa Barbara and is a visiting scientist at OpenAI. The underlying work appeared in an October preprint prepared with the aid of GPT-6 Astra.

Deep dive

The Standard Model describes the known elementary particles and three of the four fundamental forces, and it matches experiments across eighteen orders of magnitude. Craig writes that one foundational gap remains after more than fifty years: the full theory has no nonperturbative definition, meaning one that does not treat its interactions as small corrections. Lattice methods, which put spacetime on a grid, are the standard way to study the strong force like this, but giving the whole Standard Model such a definition has remained an open problem.

In the essay, Craig says he had a strategy for the problem but could not prove that it worked in every case it needed to cover. He writes that GPT-6 Astra helped turn that strategy into a construction and took the lead in developing the detailed proof. The result, proposed in his October preprint, is a candidate definition built on overlap fermions, a technique the essay traces to earlier work by Ginsparg, Wilson, Narayanan, Neuberger and Lüscher.

Craig adds that the construction leaves open how to define the interacting theory in continuous spacetime, and points to what he calls OpenAI's recent proof of the existence of Yang-Mills theory as the missing piece. He says that combination let OpenAI prove the existence of the simplest unified Yang-Mills theory that includes the gauge fields and fermions of the Standard Model. Greg Brockman shared the post as GPT-6 Astra for theoretical physics. These are claims by OpenAI and its visiting scientist, and the essay does not report independent peer review.

What to expect next

The work now needs scrutiny from lattice and mathematical physicists, starting with the October preprint and the related Yang-Mills existence claim. Watch for peer review, independent checks of the proof, and whether other groups can build on the construction. OpenAI's new science blog suggests more guest posts from researchers using GPT-6 Astra, so expect the company to keep pitching the model as a research partner and not only a coding tool.