OpenAI recently published solutions to decades-old mathematical problems that have sent shockwaves through the mathematics community. The achievement marks a remarkable acceleration in AI capability, with systems advancing from being unable to perform basic tasks like counting letters to handling sophisticated, professional-grade abstract mathematics in just six months to a year. This compression mirrors the disruption AI has inflicted on software engineering, but occurring at a far faster pace.

The breakthrough has prompted serious questioning among leading mathematicians about the future of their discipline and the utility of continuing to train generations of researchers. Key concerns include whether academic mathematics research remains valuable if AI systems can systematically solve open problems, what role university mathematics programmes should play, and whether frontier AI companies prioritise commercial applications over the field's institutional health. The paradox complicating these questions is that AI remains remarkably poor at elementary arithmetic whilst excelling at theoretical mathematics, making it difficult to predict which domains of human mathematical expertise will retain value.