I almost fully agree with you, however there are also very important "micro comment" cases. IME this comes when implementing algorithms from papers. For example, when looking at some formula in a paper, and looking at a python implementation of it that uses numpy, there may not be an obvious 1:1 mapping. Numpy has some very powerful operations, so commenting each line with how that statement maps back to the algorithm makes revisiting or others' visits much easier to comprehend.

Another case is the "reminder note"... e.g. "This function also sets up the next step..." (I know, side effects and whatnot, but sometimes you're stuck with other people's libraries).