In a guest essay on Terence Tao’s blog, Karthik C. S., Amit Sahai and Jie Xue argue that mathematical truth and understanding retain their value even when AI makes a discovery first. Longer reasoning chains or more papers do not automatically determine which questions are worth pursuing, and either using or declining AI can express a commitment to truth. A separate guest essay by Frank Calegari likewise distinguishes producing results from structural understanding. These are opinions about research culture, not experiments demonstrating AI accuracy or the effectiveness of a particular policy.