Live updates
GPT-6 Astra led the proof behind a candidate lattice definition of the Standard Model, physicist says in first post on OpenAI's new science blog
First seen on X 14 hours ago@OpenAI ♥ 1,46722 views
OpenAI launched a science blog with a guest post from physicist Nathaniel Craig, who says GPT-6 Astra drove the proof behind a candidate lattice definition of the Standard Model of particle physics. Astra "took the lead in developing the detailed proof," Craig wrote.
According to the post published Nov 15, Craig, who leads the particle theory group at UC Santa Barbara, did the work as a visiting scientist at OpenAI. The Standard Model has matched experiments for more than 50 years, yet it still lacks a nonperturbative definition, one that does not treat interactions as small corrections. The obstacle is the weak force, which treats left- and right-handed particles differently. Put a left-handed particle on a spacetime grid and an unwanted right-handed partner appears. Martin Lüscher solved the related phase problem for simpler theories in 1999, but the full Standard Model stayed open for more than 25 years. Craig had a strategy of carrying the same force fields onto ever finer grids and comparing them, but could not prove it worked in every case. He said Astra proposed how to carry the fields over and showed the comparisons converge. The preprint went up on arXiv on Oct 6.
Craig called the result "far more an answer in principle than an answer in practice" and said experts still need to vet it. Lattice theorists Maarten Golterman, Andreas Kronfeld and Yigal Shamir reviewed the proof beforehand. He tied the work to OpenAI's Oct 27 claim of a Yang-Mills existence proof. OpenAI's post on X passed 200,000 views in two hours, though the top replies read "No GPT-6.2 again" and "where new model."
“Physicist Nathaniel Craig shares how he worked with GPT-6 Astra to develop a candidate lattice definition of the Standard Model.”
