Thread

1. cantalopes · 1790031053 · focus · HN ↗

   Not a word about stealing solution to the navier stokes prize problem

   1. keeda · 1790033990 · focus · HN ↗

      They already amended their last statement to address it thus:

      > Following an investigation, we have confirmed that Buckmaster’s Codex prompts over the two months preceding this announcement and paper on September 8, 2026, could not have influenced the system in any way, including through training.

      And clearly, if they really have solved 100s of other open problems since then, it is only more reason to believe that statement.

      1. philipwhiuk · 1790077246 · focus · HN ↗

         There are hundreds of mathematicians working on problems, probably feeding data to Codex.

         1. keeda · 1790117598 · focus · HN ↗

            So, the premise is that there just happened to be hundreds of mathematicians who were close to solving these hundreds of long-standing unsolved problems with Codex all in the last few months, and OpenAI trained on all their chats to scoop them?

            Well, I do suppose the probability of that is non-zero... ;-)

Unofficial Hacker News client; not affiliated with Y Combinator.