Agree with the sentiment here as well that it applies to many "laws" of programming. There are <i>always</i> exceptions, though this one is particularly harmful. I've worked with people who don't consider <i>any</i> optimization until it starts to be a problem, even if something trivial can be done to prevent it (n+1s were an example). As engineers we seem to be constantly searching for a single answer to each common problem but that's just never going to be the case.