Comment by math_dandy

Regarding 6., that we shouldn't throw AIs at unsolved problems in bulk and thus lose human expertise. I think that ship has well and truly sailed. AI will continue to be thown at all manner of unsolved problems, and at an increasing rate - by commercial labs, by individual researchers, and by academic research teams. The mathematical community needs to establish paths to meaning, progress, and growth of human expertise in a world where AIs will by default be thrown at every unsolved problem.