@tao@mathstodon.xyz (4/4) And if we genuinely believe that insight matters more than the bare answer, then we should apply that principle consistently. We should not decide in advance that a solution is intellectually impoverished because it came from AI, nor assume that a slow human solution is inherently richer. Inspect the mathematics. Identify the structures. Extract the mechanisms. See what new mathematics becomes possible. That is how we should judge whether a solution has insight. A good problem is not a non-renewable resource. It is an entrance.