OpenAI says its AI models have produced proofs that challenge established ideas in mathematics and computer science and address a major problem in physics, Qazinform News Agency correspondent reports.
One claim challenges a long-standing belief about how quickly computers can solve difficult problems. OpenAI says its AI system disproved the Exponential Time Hypothesis, which researchers have used for decades to estimate how much computing power certain tasks require.
The hypothesis suggests that some problems become so difficult as they grow that solving them quickly is impossible. Researchers have relied on it to assess the limits of algorithms in fields ranging from biology to computer science.
If OpenAI's proof is correct, scientists may need to rethink these limits and explore faster ways to solve some problems. The finding does not mean all difficult computing problems can now be solved quickly. In mathematics, OpenAI says its models proved a 1991 conjecture about how waves behave when they focus. The result also settles two related mathematical problems, according to the company. These findings could help researchers better understand wave behavior and advance other areas of mathematics.
In physics, OpenAI says its models produced mathematical proofs concerning theories that explain how fundamental particles interact. One describes the strong force, which holds quarks together inside particles such as protons and neutrons. The work also addresses a problem involving Yang-Mills theory, one of the questions included in the Clay Mathematics Institute's Millennium Prize Problems.
OpenAI has published the mathematics and computer science proofs for examination. Independent experts will review the computer science claim, while a separate team led by Nobel Prize-winning physicist David Gross will examine the physics findings.
Earlier, Qazinform News Agency reported that OpenAI expanded ChatGPT ads to 26 more countries.