Abstract page for arXiv paper 2608.16977: The Problem Is the Problem: Towards Scalable Mathematical Discovery