Claude team - Love what you guys have built, but enforcing adaptive thinking should be rethought. Barring trivial cases, what needs more attention from the model vs not is contingent upon the user's expectations of the answer. Absent deep contextual knowledge of what a user prefers, this is unlikely to yield satisfactory answers. I have been a Max subscriber for a long time and as far as I can recall, I go the extra mile to make sure Claude has the right level of nuance, independent of what the model feels is optimal. Hope that you'd revisit this change. @bcherny@claudeai@trq212@felixrieseberg