OpenAI claims disproof of Exponential Time Hypothesis in three-field breakthrough

Related stories

• OpenAI releases 722 AI-generated math manuscripts, claims on quasi-Riemann and Kakeya

  OpenAI published 722 mathematical manuscripts produced by an unreleased internal frontier model, grouped into 372 result families across 17 fields, including claims on the quasi-Riemann hypothesis, BSD conjecture, and four-dimensional Kakeya conjecture. The company said most results came from a single prompt to a single AI agent, with average compute per result roughly equivalent to three hours of ChatGPT Pro thinking. Mathematicians described the release as overwhelming, noting it could take years to fully understand, and OpenAI withdrew three papers two days later.

  16 reports · 6 sources1334 points · 1520 comments3 weeks ago
  

• OpenAI releases 700 AI-proved math results; cryptography problems absent, sparking speculation

  OpenAI has released more than 700 mathematics results proved by its latest model, a development commentators describe as ending human mathematics as known. Observers noted that none of the results touch modern cryptography, prompting speculation that the omission is deliberate to avoid giving competitors an asymmetric commercial advantage. Cryptographer Matthew Green commented that public key cryptography might be lost.

  1 reports · 1 sources54 points · 19 commentslast week

• Fired OpenAI safety researchers urge halt to AI monitoring work, dispute misconduct claims

  Three fired OpenAI safety researchers—Jasmine Wang, Tomek Korbak, and Mikita Balesni—published an open letter urging AI labs to halt work that could impair AI monitoring, saying their dismissals are chilling remaining staff. OpenAI defended the firings, citing an investigation that found they violated policies on handling sensitive information, while the researchers dispute the claims and say they were fired for prioritizing safety over company interests.

  7 reports · 5 sources365 points · 228 comments3 weeks ago
  

• OpenAI withdraws 3 math papers from openai/math repo over sign error, revises 14

  OpenAI has withdrawn three AI math papers from its openai/math repository after discovering a sign error in a key proof, reducing the total manuscript count from 722 to 719. The withdrawn papers, all related to the Hodge conjecture, relied on a construction where a geometric operation was incorrectly signed as +1 instead of -1, invalidating the proof. OpenAI also revised 14 other papers and updated citations in 13, while emphasizing that the withdrawals concern the proofs, not the mathematical statements themselves.

  1 reports · 1 sources3 weeks ago