Mathematician Tristan Buckmaster stated that OpenAI exerted pressure on him after details leaked about his progress in researching Navier-Stokes equations using AI. According to him, the company attempted to remove a co-author employed by Anthropic and threatened him after he refused.
The Decoder reports that OpenAI subsequently announced its own breakthrough, utilizing the same unusual solution path. Buckmaster had previously uploaded all his drafts to Codex. He states that OpenAI claimed the system did not access user data but did not answer his question regarding whether the materials were used for training.
The significance of the story relates to authorship, the confidentiality of research materials, and the transparency of AI training. However, the editorial team only has metadata and a retelling from The Decoder, without confirmation from OpenAI, Anthropic, or independent sources.
editorial commentary
Why it matters
A likely consequence is strengthened requirements for authorship rules, access to user drafts, and disclosure of AI training practices. The next verifiable signal will be OpenAI's public position and the availability of documents or independent confirmations. The main uncertainty remains the absence of primary materials and responses from the parties in the available source.