The authors of the Exponential Time Hypothesis (ETH) are now the people checking whether two AI labs knocked it down. Russell Impagliazzo and Mohan Paturi of UC San Diego, working with MIT’s Ryan Williams, are coordinating outside scrutiny of claims by OpenAI and Anthropic that their models disproved ETH, UC San Diego Today reported on November 19.

Neither lab’s result has been confirmed. The group opened a blog, ETH: Community Examination and Discussion, where researchers can post findings, and it has published the OpenAI prompt behind the claim. It also plans one or more academic conferences to discuss verification work. Impagliazzo and Paturi proposed ETH in 1998 with then-PhD student Francis Zane.

UC San Diego says that if the findings hold, the consequences could reach chip design and internet encryption. The university’s account offers no verdict on the proofs themselves. Until reviewers report back, treat any system built on ETH-based hardness assumptions as unchanged, and watch the blog for the first independent checks.

UC San Diego Today, November 19, 2026.