OpenAI, Anthropic Claim AI Models Refuted Exponential Time Hypothesis; MIT's Williams Leads Review

6.0 · 5d ago · WORLD · AI

The Facts

  • OpenAI announced that its AI models made three scientific discoveries in harmonic analysis, complexity theory and quantum field theory. ✓2 · Al Jazeera, TechCrunch
  • OpenAI says its models disproved the Exponential Time Hypothesis (ETH), an assumption about the time needed to solve hard computational problems. ✓2 · Al Jazeera, TechCrunch
  • OpenAI says its models proved the local smoothing conjecture, a central proposition in harmonic analysis, in every spatial dimension for the first time; researchers had previously proved it only in three dimensions. Al Jazeera only
  • OpenAI says its models provided a mathematically rigorous proof of the existence of quantum chromodynamics, the theory of quark and gluon interactions. Al Jazeera only
  • OpenAI used a different release approach for each of three fields: mathematicians got the work on GitHub; computer scientists formed a working group, asked for the proofs to be shared publicly and for OpenAI to fund a conference; physicists formed a working group that will decide how their result is released. TechCrunch only
Dive deeper · 8 more factsShow less
  • Al Jazeera reports OpenAI consulted three expert groups as 'shepherds'; the mathematics and computer science groups opted to publish the AI-produced data immediately, while the physics group deferred release to write its own interpretation. ⚠ disputed
  • Anthropic released its own refutation of ETH hours after Anthropic mathematician Levent Alpöge suggested OpenAI's solutions were based on pre-existing work. TechCrunch only
  • The claimed algorithms solve the relevant logic puzzles in 'sub-exponential' time, faster than previously thought possible. ETH was first published in 1999 by Russell Impagliazzo and Ramamohan Paturi of UC San Diego. TechCrunch only
  • MIT professor Ryan Williams is organizing the computer science community's response, with Impagliazzo and Paturi helping interpret the solution. The group released an open letter seeking feedback and discussion of norms for AI contributions. TechCrunch only
  • Williams said the proof has routine parts and 'two or three alien-looking parts', that many papers relying on ETH would need revisiting if the claim holds, and that improved algorithms could raise cryptographic concerns for certain key sizes. He said he doubts OpenAI's algorithm is the best possible. TechCrunch only
  • Alex Iosevich, a University of Rochester harmonic analysis professor, said OpenAI appeared to have 'cracked' the local smoothing conjecture, calling it a 'historic development'. Al Jazeera only
  • OpenAI acknowledged tradeoffs in its release approach, saying the choice of initial shepherds shapes who gets involved first and creates information asymmetry, and invited feedback. Al Jazeera only
  • Williams had earlier this October called an OpenAI proof dump a 'DDOS attack on mathematicians'. TechCrunch only

Context

OpenAI says its AI models produced solutions to three major problems in mathematics, computer science and physics, including a refutation of the Exponential Time Hypothesis, a widely used assumption underpinning understanding of computational hardness. Anthropic has since released its own claimed refutation of the hypothesis.

The claims have not been independently validated. A working group led by MIT's Ryan Williams is reviewing the computer science result, and OpenAI has been asked to fund a conference to discuss it.

Conflicting Reports

  • Al Jazeera says the computer science group opted to publish the AI-produced data immediately; TechCrunch says the computer scientists formed a working group and asked for the proofs to be shared publicly and for OpenAI to fund a conference, with no mention of immediate publication.

How the Agent Panel Scored This

Independent AI personas each scored importance 1–10; the story's score is a trimmed mean (highest and lowest votes dropped). Vote spread: 3.

7 Global Affairs Analyst AI systems claimed to refute the Exponential Time Hypothesis and prove major conjectures, with Anthropic issuing a competing refutation, which would have broad consequences for computer science and cryptography if validated, though the claims remain unverified.
7 Science & Tech Analyst If validated, AI-produced refutation of the Exponential Time Hypothesis plus claimed proofs of local smoothing and QCD existence would be a major scientific milestone, though the claims remain unverified and the story reports announcements and reviews rather than confirmed results.
6 Business Analyst Unverified AI-generated claims of refuting a foundational complexity hypothesis and proving local smoothing in all dimensions are significant research news with potential cryptographic implications, but they remain unvalidated and mainly affect academic communities, so moderate-high importance is warranted.
6 News Editor The story reports concrete claimed results (disproof of ETH, a proof of local smoothing conjecture in all dimensions, and a claimed QCD existence proof) with potential field-wide consequences, but the claims are unvalidated and the impact is largely confined to research communities rather than the general public.
6 Critical Reader The story reports concrete but unverified AI-generated claims (ETH refutation, local smoothing proof, QCD existence proof) plus a rival Anthropic refutation, which is notable for theoretical computer science and math, but the results are not validated and impact is mostly within research communities.
5 Policy Expert The story reports unverified corporate claims of AI-generated proofs and a rival refutation, with concrete actions limited to research releases and working groups rather than enacted laws, rulings, or widespread immediate impact, though the potential implications for cryptography and computational theory make it notable.
4 Informed Citizen The story reports unverified claims of AI-produced proofs in specialized theoretical fields with no immediate impact on ordinary people, though the potential cryptographic implications and the disputed nature of the claims give it some notability beyond niche academic interest.

Sources (2)

← All stories