OpenAI published solutions to longstanding math problems, sparking debate among mathematicians about the field's future. The Verge's Robert Hart reports that AI models remain poor at basic arithmetic but have improved rapidly at abstract math. Mathematicians are questioning their role if frontier models can solve outstanding problems, while some speculate the attention around math serves as marketing for AI labs. Hart spoke with leading mathematicians about this "existential crisis."
No score is assigned. Sources and their independence are shown in the citation chain below.