This is a phenomenal general rule and can be expanded well beyond engineering and social norms:

At a certain point it just doesn't make sense to over-optimize for productivity, when the sacrifice {insert impact} isn't worth the gain {insert output}.