My rough rule is that I’d like to write stuff that will still be worth reading in five years, and ideally stuff that will be more relevant a year from now. Because of the way news site algorithms currently work, that’s the opposite of what everybody who writes for a living does.