Home / Science / AI Solves 10 Math Mysteries, Shakes Up Field
AI Solves 10 Math Mysteries, Shakes Up Field
3 Aug
Summary
- AI model Astra from OpenAI solved ten complex mathematical problems.
- Soficity counterexample found, disproving a long-held mathematical conjecture.
- Mathematicians express concern over AI's reliance on prior human work.
OpenAI's AI model, Astra, has successfully solved ten significant mathematical problems that have puzzled experts for years. Among these breakthroughs is a crucial counterexample to the conjecture of soficity, a property related to approximating infinite group operations. This discovery challenges the long-held belief that all countable groups possess this property.
Mathematicians are reacting to this rapid advancement, with some, like Francesco Fournier-Facio, raising important questions about AI's independence. He suggests OpenAI's solutions, particularly to the soficity problem, heavily build upon prior human research, advocating for greater recognition of that foundational work. He notes that many recent AI mathematical successes focus on finding verifiable counterexamples rather than developing novel theories.
Abhishek Saha acknowledges the impressive nature of Astra's solutions, including proofs beyond just counterexamples. He highlights that these achievements, while technically complex, largely involve assembling existing mathematical concepts in new ways. Saha anticipates that AI models may soon progress to developing entirely new mathematical theories within the next few years.
