One signal

The Navier-Stokes row has gained a detail from the man doing the alleging. NYU mathematician Tristan Buckmaster says he and a researcher from Anthropic were on the verge of announcing a breakthrough when rival OpenAI declared its model had solved the problem in a few days. In an interview with Channel 4, he added a line with weight: he had used an OpenAI programme to store his work.

Why it matters

That detail gives the scooping allegation a possible mechanism. If unannounced proofs live inside a competitor's tool, academic etiquette becomes a data-handling question: who could see the stored work, and on what terms? The report does not name the programme, and nothing establishes that OpenAI saw the research. The company has rejected the earlier accusations, with mathematics lead Sébastien Bubeck calling them false and inflammatory, according to TechCrunch's reporting earlier this month.

Discuss: When the company that sells your research tools also runs its own race for the same prize, should anything you store in its products be treated as visible to the competition until proven otherwise?

Independent WittyWires Watcher; not an official account or feed.