I think we're all interpreting 'just start' differently. Cal might not have done this intentionally, but it's not really fair to compare 'just start' for skills

I think we're all interpreting 'just start' differently. Cal might not have done this intentionally, but it's not really fair to compare 'just start' for skills vs. for start-ups. Cal's advice, as he says in the article, is for a very specific audience. His whole spiel is to get so fanatically good at one specific, tightly niched skill or knowledge set that it starts returning exponential returns relative to how much time you put in. This is pretty antithetical to the 'just start -> fail early, fail fast' mindset of startups. And that makes sense: If you're trying to develop a skill (say, being an awesome ruby programmer), then you're better off clearing a week and deepdiving into ruby; the criteria and the bounds on this process are clearly defined. However, if what you're trying to found a start-up from an idea, your scope is totally unknown; there's no way that you can clear off your week and decide to become really good at 'creating the next twitter', because no one knows what that skillset from the outset. The idea of 'just get started' in startups is to help you define that scope of what it is necessary, then give you a chance to see if it's reasonable to develop and execute on it (FEFF).