OpenAI and Anthropic claim models refute the Exponential Time Hypothesis

Published on Dorsal: October 30, 2026 at 8:52 AM (UTC)

Editorial update: October 30, 2026 at 10:35 AM (UTC)

Source date: October 27, 2026 at 5:08 PM (UTC)

AI-generated conceptual illustration for “OpenAI and Anthropic claim models refute the Exponential Time Hypothesis”. — AI-generated illustration

Frontier labs released solutions to long-standing math problems and claim their models produced algorithms that run in sub-exponential time, prompting working groups and public scrutiny from computer scientists led by MIT’s Ryan Williams over technical and cybersecurity implications.

Summary

AI-assisted summary of the linked source

TechCrunch reports that OpenAI published solutions to three high-profile math problems — including a claimed refutation of the Exponential Time Hypothesis (ETH) — and that Anthropic released a separate refutation after initial criticism; MIT professor Ryan Williams is coordinating a field response and has urged community review while probing whether the results affect cryptographic systems, which OpenAI says it checked and found reasonably confident did not imply breaks.

Go to source