OpenAI and Anthropic Say Models Refuted ETH, Impagliazzo and Paturi Open Review Blog
Summary
OpenAI and Anthropic say their AI models refuted the Exponential Time Hypothesis, a 1998 conjecture by UC San Diego's Russell Impagliazzo and Mohan Paturi. The two, with MIT's Ryan Williams, launched a blog, ETH: Community Examination and Discussion, to verify the claim independently and posted the OpenAI prompt behind it.
Key Points
- Paturi also serves as associate dean for artificial intelligence strategy at the UC San Diego Jacobs School of Engineering.
- The team plans academic conferences where verification work and farther-reaching implications will be discussed.
- ETH was proposed in 1998 with then-PhD student Francis Zane, now an independent researcher.