OpenAI and Anthropic claim AI models disproved key theorem
The Exponential Time Hypothesis sets a minimum difficulty for solving certain problems, and academics are now questioning the companies' claimed disproof.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
This story is no longer being updated. It reflects reporting up to .
What happened
OpenAI and Anthropic announced Monday that AI models built in-house had disproven the Exponential Time Hypothesis, a long-held computer science result. Academics criticized the release of unvetted AI-generated proofs and questioned who would be accountable for checking them. Some researchers said the work has made decades of prior research seem irrelevant, while others described feeling eager to explore new questions.
Why it matters
Some researchers say decades of related work may now seem irrelevant.
Timeline
- Monday: OpenAI and Anthropic say their internal AI models disproved the Exponential Time Hypothesis.
- Oct 28, 11:37 AM ET: Academics criticize the release of unvetted AI math proofs and question who will review them.
- Oct 29, 6:45 AM ET: Researchers say they are still adjusting to the result while looking at new questions to explore.
Coverage
- Encryption Under Siege? OpenAI, Anthropic Disprove Key Cryptography Theorem (Jemima McEvoy)
- An AI Breakthrough Has Made Decades of Computer Science Research ‘Irrelevant’ (wired.com)
- OpenAI Solved Hundreds of Math Problems. Then It Left Mathematicians to Clean Up the Mess. (vocal.media)
- OpenAI Proved Half of Yang-Mills. The Missing Half Is the One Pauli Would Not Drop. (vocal.media)
- Who Will Review OpenAI's AI-Generated Math Proofs? (Koo A-mo)
- OpenAI Says Its AI Refuted the Exponential Time Hypothesis. Is Bitcoin Safe? (BeInCrypto)
- OpenAI Advances Proactive Agents and AI Math Breakthroughs While Facing Lawsuit Over ChatGPT Medical Advice (theaiinsider.tech)
More from Technology & AI
- China puts humanoid robots to work serving food and sorting parcels
- Newsom proposes sweeping AI safety rules for Washington
- Google releases Gemini 4 Neon AI model with million-token context
- Microsoft and Intel widen their cloud chip partnership
- Google DeepMind rolls out Genie 4 to select subscribers
- AI helped AMD shorten chip design work, OpenAI says