These days I find myself more #agreeable to the #measurement #argument - rather than jumping straight into #implementation https://danluu.com/why-benchmark/ The thing you need to be able to answer is "is this new thing actually better" or "will doing this new thing even have a chance of being better"?