AI is solving 'impossible' math problems. Can it best the world's top mathematicians?

Meta and Google's AI systems have tackled century-old mathematical problems, with Meta's model solving stability equations at a 10.1% success rate and Google's Gemini achieving gold-medal performance on high school math olympiad problems. However, mathematicians remain divided on whether AI represents true breakthrough thinking, as current systems still require significant human guidance and haven't yet proven theorems beyond human capability, though experts predict AI may soon solve thousands of mathematical conjectures simultaneously.