Deciding before the picture is complete
Most of the pain in building isn't the work itself — it's the hovering before the work starts.
Yesterday I made a dozen small decisions without stopping to agonize over them. Swap out a system that wasn't working. Change how something routes. Restructure a flow before adding more to it. None of them were obvious in advance. None of them came with proof they were right.
I used to think decisiveness was a personality trait — something you either had or didn't. Now I think it's more like a muscle that gets built through repetition, and the repetition happens whether you want it to or not when you're running something alone.
The recursive thing about solo work is that every decision is load-bearing. There's no committee to absorb the consequences. If you choose wrong, you fix it. If you hover too long, the hovering itself becomes the problem — the backlog grows, the context degrades, and you arrive at the decision in worse shape than when you first faced it.
The cost isn't the wrong decision
I've come to believe the expensive move is almost never choosing wrong. It's the delay. Wrong decisions teach you something and can usually be unwound. The delay teaches you nothing and compounds quietly — more code written on top of a shaky assumption, more customers onboarded into a broken flow, more days spent maintaining something you already knew wasn't right.
Naval's idea that decisiveness is the most important skill you can develop used to feel like a productivity bromide to me. It doesn't anymore. What he's pointing at is something specific: when information is incomplete — and it is always incomplete — the founder who acts has an advantage over the founder who waits for certainty that will never arrive. You don't get more information by hovering. You get more anxiety.
This doesn't mean being reckless. There's a version of fast decision-making that's just avoidance — picking quickly so you don't have to sit with the discomfort. That's not it. What I'm describing is something quieter: building enough pattern recognition over months and years that you can read a situation fast, make a call, and move.
It took me a long time to realize that the goal isn't to make the perfect call. The goal is to keep the work moving and stay close enough to the output that you can catch the mistakes before they compound too far.
Most of the decisions you're hovering over right now don't need more time. They need you to pick.