OpenAI and Anthropic claim models disproved key computing hypothesis

OpenAI and Anthropic say their models have disproven the exponential time hypothesis, which underpins a substantial body of theoretical computer science. The companies’ claims point to a potentially significant challenge to longstanding research, with decades of work described as having been rendered irrelevant. That sweeping characterization remains distinct from the narrower reported claim: that the models have disproven a hypothesis central to the field.

Researchers’ reactions include reports of emotional distress. A department has considered hiring an occupational psychologist because “some people are really depressed and really sad.” The reported response highlights the personal stakes surrounding claims that AI systems have overturned an established theoretical foundation, alongside the implications for the relevance of research built around it.