> don't prematurely optimize
Premature optimization is the root of all evil
Premature optimization is the root of all evil
You may end up using them as arguments when they're just catchphrases that require context.
See people who used "goto considered harmful" as an absolute to never use goto (C).
Seriously true, though. What's usually missed is the first bit, about "small efficiencies", combined with the last bit, the "critical 3%". He's basically saying don't spend time in the mass of weeds until you know where the problems really are, but conversely take the time to optimise the smaller areas that are performance-critical.
Almost the opposite of the carte-blanche overly-minimal MVPs often thrown together by those of us who forget what the V stands for.