OpenAI says its models produced three new results, including a disproof of the Exponential Time Hypothesis
Outside experts acting as "shepherds" recommended a different release plan for each of the math, computer science and physics results.
On October 26, OpenAI announced that its internal models had produced three research results in mathematics, computer science and physics. One is a claimed disproof of the Exponential Time Hypothesis, a long-standing conjecture in computational complexity theory. OpenAI calls the AI-generated manuscripts raw research data that still has to be turned into scientific understanding. The company asked outside experts to act as "shepherds" and organize groups to work through the results. The shepherds recommended three approaches. For mathematics, OpenAI will release the AI output directly and leave researchers to organize their own explanatory work. For computer science, it will form an outside group of collaborators and also publish the output for anyone to examine. For physics, it will form a group and defer to that group on what to release. OpenAI says it does not yet know which approach will work best and has asked for feedback. The collaborations will follow four principles, including transparency with no nondisclosure agreements. The Verge notes that the announcement comes less than a month after OpenAI released hundreds of manuscripts it said solved long-standing mathematics problems. The claimed results still need independent verification, and a confirmed disproof of the hypothesis would be a significant development in complexity theory.