Automation Without Understanding: The Strategic Error of Losing Math Capacity

Automation Without Understanding

I argue that relying on AI for research while weakening the human pipeline to understand it is a critical strategic error. Mathematical capacity is vital infrastructure, not just a byproduct of theorem production. Drawing on the recent AI disproof of an Erdős conjecture, I propose treating this human skill like semiconductor capability and mandating that AI systems expose their reasoning in formal, machine-checkable forms.

"Mathematical capacity, which is the trained ability to verify, interpret, and challenge mathematical reasoning, is not a byproduct of theorem production but a form of infrastructure, built over generations by institutions that cannot be reconstituted on demand."