🚨 OpenAI says its AI has DISPROVED ETH. ETH here stands for the Exponential Time Hypothesis, a decades-old assumption about how difficult certain computing problems are to solve as they grow. OpenAI says its AI developed an algorithm faster than the hypothesis permits, challenging an assumption used in hundreds of research papers. The company says the disproof was found on October 16, verified the next day using the Lean proof-checking system, and published on GitHub alongside its formal verification. ETH’s original proposers, Russell Impagliazzo and Ramamohan Paturi, will lead a review with MIT’s Ryan Williams. Turning the result into faster real-world algorithms “will require significant further research,” OpenAI says. The company also claims two further breakthroughs: • Mathematics: Resolving the Kakeya conjecture in every dimension, following last year’s proof of the three-dimensional case. • Physics: Producing a proof addressing part of a $1 MILLION Millennium Prize problem. That work remains unreleased, with a group led by Nobel laureate David Gross set to examine it.