OpenAI and Anthropic claim models refuted the Exponential Time Hypothesis
Original title: Frontier AI labs say their models refuted the Exponential Time Hypothesis. What does that mean?
Computer scientists are reviewing claimed sub-exponential algorithms, including whether they weaken assumptions behind some cryptography.
OpenAI released claimed solutions to three long-standing math problems, including a rejection of the Exponential Time Hypothesis. Anthropic later released its own claimed refutation after a company mathematician suggested OpenAI drew on existing work. MIT’s Ryan Williams is organizing a computer-science response and said OpenAI reported it was reasonably confident the result does not break cryptographic systems.
Unconfirmed: The labs’ claims are not independently confirmed, and Williams said OpenAI’s algorithm is probably not the best possible.
- OpenAI published the math work on GitHub and is working with computer scientists and physicists separately.
- ETH was first published in 1999 by Russell Impagliazzo and Ramamohan Paturi.
- Williams said two or three parts of OpenAI’s proof look unusual and need community review.
- He said the result, if it holds, would force many papers that rely on ETH to be revisited.
- Organizations
- OpenAIAnthropicMITUC San Diego
From the original article
4 min read at techcrunch.com
TechCrunch spoke to the MIT computer scientist charged with explaining OpenAI's latest mind-bending discovery
By Tim Fernholz · techcrunch.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.