The 722 preprints released on GitHub have deeply divided the mathematics community, shifting focus from the results to the ethics of AI research. Fields medalist Hugo Duminil-Copin stated the dump "annihilated" his specific field of research, while Terence Tao warned that junior research projects were wiped out overnight. Conversely, ex-OpenAI researcher Scott Aaronson argued that published problems belong to the whole world, and other experts view the AI's ability to interpret incomplete research notes as a significant advancement in research tools.
Story
NYU Professor Accuses OpenAI of Strip Mining User Data in Math Release
AI5 days ago
Tristan Buckmaster of New York University claims an internal OpenAI model was used to harvest private grant applications and research notes to fuel a public display of technical dominance. In a guest post on Terence Tao's blog, Buckmaster describes the October 6 release of over 700 mathematical papers as a "PR stunt" that competed against its own users, some of whom were days away from publishing their own work. He also challenged the AI lab's claim of three hours of computation per result, arguing the figure is meaningless because it ignores failed attempts and prior agent work.
Looking for more context or analysis?
Tilt chat provides reliable, tailored insights, alerts, and automations.