🔍 Read the full analysis: Why OpenAI’s AI Claim About The Exponential Time Hypothesis Needs Context on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
OpenAI says an unreleased AI model produced a proof that refutes the Exponential Time Hypothesis, a long-standing assumption about the difficulty of solving certain logic problems. A Lean formalization checks a weaker runtime bound than the paper’s sharpest result, and independent expert review has not yet concluded. The claim does not resolve P vs. NP or show that practical computing will get faster soon.
The claim concerns k-SAT, a family of problems that asks whether truth values can be assigned to variables so that every rule in a formula is satisfied. For 3-SAT, ETH holds that no algorithm can solve every instance in time that grows subexponentially with the number of variables. The preprint describes a deterministic algorithm that exactly counts satisfying assignments for fixed k in subexponential time and polynomial memory. If correct, its 3-SAT result contradicts ETH.
OpenAI’s October 26 announcement said its system found the disproof on October 16, with the Lean formalization following a day later. The preprint is dated October 17 and was revised on October 25. The source article reports that the public Lean release contains 3,977 files and about 710,000 lines, while the formal statement itself is short. The article’s author says they scanned the code; that is not an independent expert verification.
The distinction between the paper and the formalization matters: according to the source, Lean checks a slightly weaker bound than the preprint’s sharpest runtime claim. The preprint also says the result refutes the Strong ETH and three related hypotheses. Those are claims in a research manuscript, not conclusions that have received independent sign-off. OpenAI says an outside group led by Russell Impagliazzo, Ramamohan Paturi and Ryan Williams will examine the computer-science result.
THREE RESULTS, THREE RELEASE STRATEGIES
What OpenAI released, what Lean checked, who reviews it
Dates and the physics status are OpenAI's own account. Lean-checked means a machine verified the proof, not that experts have accepted it.
Local smoothing, so Kakeya too
- AI output
- 18 Oct 2026
- Lean
- 19 Oct 2026 · strict ranges
- Public
- AI write-up + Lean proof
- Review
- No group: researchers organize themselves
The Exponential Time Hypothesis is false
- AI output
- 16 Oct 2026
- Lean
- 17 Oct 2026 · weaker bound
- Public
- 67-page preprint + Lean proof
- Review
- Impagliazzo, Paturi (UC San Diego), Williams (MIT)
QCD exists; pure Yang–Mills as sub-result
- AI output
- 9 Sep (Yang–Mills), 18 Sep (QCD)
- Lean
- Yang–Mills existence, 16 Oct (OpenAI)
- Public
- No: the group holds the AI materials back
- Review
- Gross, Douglas, Jaffe, Klebanov, Nekrasov
Before this release
The repository the results sit in
- 6 Oct: more than 700 AI-generated manuscripts published
- 7 Oct: 3 withdrawals after a sign error; 14 other manuscripts revised
- About 42% of top-line results formalized in Lean (repository’s own figure)
- OpenAI calls the manuscripts “raw research data”, not polished papers
Sources: OpenAI, 26 Oct 2026 · Lean note 378 · Lean note 079 · repository history. Status chips: our assessment.
Powered by Thorsten Meyer AI
A Major Test for Complexity Theory
ETH has served as a foundation for many conditional hardness results: researchers use it to argue that particular problems are unlikely to have substantially faster algorithms. If the proof survives expert review, those arguments may need to be revisited, and some established expectations about algorithmic limits would change.
The immediate significance is also methodological. The episode tests a workflow in which an AI system proposes a proof and a proof assistant checks a formalized version. Lean’s verification can establish that the formal code follows its encoded definitions and logic; it does not by itself show that the code captures every important claim in the paper. Here, the gap between the formalized bound and the paper’s sharpest bound makes human review especially relevant.
There is no announced near-term practical speed-up. The stated algorithm remains superpolynomial, and the source says OpenAI itself expects practical improvements to require significant further research. The result, if validated, changes a theoretical belief, not the performance of everyday software overnight.
Products Worth Considering
Samsung 32-Inch Odyssey OLED G8 (G80SH) 4K Gaming Monitor
LG UltraGear 32GX850A-B 32-Inch 4K UHD Glossy OLED Gaming Monitor
Apple iPad Air (4th Generation, 10.9-inch, Wi-Fi, 256GB) – Space Gray
As an affiliate, we earn on qualifying purchases.
What ETH Says About 3-SAT
In a k-SAT formula, each rule is a logical clause involving no more than k variables or their negations. The task is to find an assignment that satisfies all clauses, or determine that none exists. With n variables, checking every possible assignment takes 2n possibilities. Researchers have developed faster-than-brute-force algorithms, but ETH says the exponential dependence cannot be eliminated for 3-SAT.
Impagliazzo, Paturi and Francis Zane formulated ETH in 2001. It is stronger than the familiar P vs. NP question, which asks whether every problem whose solution can be checked efficiently can also be solved efficiently. A refutation of ETH would not settle P vs. NP: the claimed algorithm is still slower than polynomial time, and P vs. NP remains open.
The OpenAI announcement covered two other research claims as well: a mathematics result concerning local smoothing and related conjectures, and a physics claim about quantum chromodynamics and a grand unified completion of the Standard Model. The source says the materials for those claims are being handled differently by outside academic groups. The ETH preprint and Lean code were released publicly at the computer-science group’s request.
““there’s by no means a consensus in the field that ETH is true””
— Scott Aaronson, in a 2017 survey, as quoted by ThorstenMeyerAI.com
Review Gaps and Boundaries
The central open issue is whether independent experts will validate the proof and whether the paper’s full runtime claim can be matched to a formal Lean proof. The public formalization covers a weaker bound, according to the source, and no expert group has yet publicly signed off on the result.
It is also not yet clear what revisions, if any, the reviewers will request, how long their examination will take, or whether the announced refutations of SETH and the related hypotheses will be accepted alongside the ETH claim. The source describes the review group and release plan but does not report a completed assessment.
The strength of the broader claim about ETH’s standing also requires care. OpenAI presents ETH as a widely held assumption, while the source cites published commentary questioning whether there is consensus. The result’s importance does not depend on describing the hypothesis as universally accepted; its use in theoretical computer science is the relevant point.
Independent Review Comes Next
The next milestone is the examination by the outside group led by Impagliazzo, Paturi and Williams. Impagliazzo and Paturi are two of ETH’s original authors, according to the source. Their review is expected to address the proof’s mathematical argument and the relationship between the paper and its formalized version.
Readers should look for a public assessment, any corrections to the preprint or Lean code, and clarification of which runtime bounds have been formally checked. Until that work is complete, the appropriate description is that OpenAI has announced a claimed refutation, with a machine-checked formalization of a weaker bound—not that the field has conclusively discarded ETH.
Key Questions
What is the Exponential Time Hypothesis?
ETH says that 3-SAT cannot be solved in subexponential time in the number of variables. It is a hypothesis used to support conditional hardness results in theoretical computer science.
What does OpenAI say its AI proved?
OpenAI says an unreleased model produced a deterministic subexponential-time algorithm for exactly counting satisfying assignments for fixed-k SAT formulas. For 3-SAT, that would contradict ETH if the proof is correct.
Has the proof been independently verified?
Not yet, according to the source. A Lean formalization is public, but an outside group is due to examine the result. The formalization checks a weaker runtime bound than the paper’s sharpest claim.
Does this solve P vs. NP?
No. The claimed algorithm remains slower than polynomial time. A refutation of ETH would weaken a stronger assumption about computational hardness, but it would not settle the open P vs. NP problem.
Will computers become faster because of this result?
No immediate practical improvement has been established. The source says the algorithm’s speed-up grows very slowly with problem size and that practical gains would require significant further research.
Source: ThorstenMeyerAI.com
