[JMostert]

You're kidding, right? Obsessing over (potential) micro-optimizations is a timeless programmer pastime. It will never go away.

Imagine something worse: suppose it had made a difference, in the current iteration of the compiler and the framework. Can you imagine the damage done by people who will now spread the word that "you must write the empty string this way and not that way", which would then continue to waste time forever, even after it's been rendered irrelevant?

This issue always arises when a language gives you two (or more) ways of accomplishing the same thing: naturally people will want to know which is the best way. Best in what aspect? Well, performance is the easiest one to objectively evaluate, so let's go with that. Cue an endless stream of benchmarks and criticism of said benchmarks, and very little reflection on how much it actually matters in terms of both programmer and user productivity.