Gary Marcus posted screenshots of Terence Tao's thread on X. Tao noted that pure mathematics problems are typically posed because efforts to solve them have historically driven further development in the field, not because the solutions themselves are urgently needed. Replies from other users discussed whether AI models should emphasize brute force or new elegant tools, how many fresh theorems might emerge, and concerns about the United States slowing AI work while competitors advance. The posts highlight ongoing conversation around Tao's perspective on directing AI toward fresh discoveries.