OpenAI has announced three new discoveries it describes as “major” across mathematics, computer science, and physics. This follows its earlier release, less than a month prior, of hundreds of manuscripts said to contain solutions to long-standing mathematics questions.
According to the description, what distinguishes this announcement is the involvement of external experts, which OpenAI calls “shepherds,” to guide the results in computer science and physics. For the mathematics problem, OpenAI says these advisers recommended releasing the AI output so that researchers could work through their own explanations. The results were added to GitHub, including work described as “Local smoothing for the Euclidean wave equation,” with a proof said to use an auxiliary parameter.
Why it matters
The shift toward involving external experts to vet or guide AI-produced results reflects a new approach to how such outputs are reviewed and shared with the research community.
Who should care
Researchers in mathematics, computer science, and physics, as well as those following how AI systems contribute to scientific work, may find the published results and the new collaboration model relevant.