OpenAI claims major proofs in math, computing and quantum physics
OpenAI said outside experts will review the claims as the company releases math proofs and funds independent work on the findings.
- On Monday, OpenAI announced an internal model produced a claimed refutation of the Exponential Time Hypothesis, releasing the model's write-up and a Lean-verified proof on GitHub.
- To manage scrutiny, OpenAI enlisted external experts to act as "shepherds" guiding the results. Russell Impagliazzo, Ramamohan Paturi, and Ryan Williams lead an independent group examining the ETH claim.
- Beyond the ETH refutation, the model produced a proof of the local smoothing conjecture in every dimension, resolving the Kakeya conjecture. OpenAI also claims a rigorous proof for certain interacting quantum field theories in four dimensions.
- The AGMAI panel of math advisers advised yesterday against designating specific people to receive output first, citing "unequal access, borrowed reputation, and biased selection" as primary concerns.
- OpenAI committed $5m to the Simons Foundation for workshops on AI-produced results. Researchers are debating how credit should recognize AI discoveries versus human verification work.
55 Articles
55 Articles
The company attributes to its models three results in mathematics, computing and physics, including half of one of the Millennium Problems, and has asked external experts to check them out.
ChatGPT developer says AI models produced proofs addressing longstanding scientific problems, seeks independent academic scrutiny
Exponential Time Hypothesis: AI models have produced proofs challenging major theories in maths, computing, physics: OpenAI
OpenAI has claimed its AI models have disproved the Exponential Time Hypothesis, which impacts computer science research. The findings include proofs related to harmonic analysis and quantum field theory, influencing various scientific domains. Researchers will verify these claims, and OpenAI plans to share the results for independent examination. The company is also funding a conference to discuss the implications of these discoveries.
Coverage Details
Bias Distribution
- 37% of the sources lean Left, 36% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
