How do you make decisions when you have incomplete information?

I pick the better option and move.

When I started building for the agency founder, I didn't know Rails. I didn't know healthcare. I didn't know HIPAA. I didn't know ABA therapy. I didn't know how New York Early Intervention agencies work. I had incomplete information across every dimension, and I had users waiting.

So I made opinionated decisions early. Next.js and Convex for Tiny Time Keeper because I needed real-time data and could ship fast. Rails for Tiny Thoughts because the domain was complex enough to benefit from convention over configuration. These weren't perfect decisions. They were good enough decisions made quickly, with the understanding that I'd course-correct once I had better information.

The philosophy comes from jiu-jitsu. On the mat, you never have complete information. You feel pressure, you read your opponent's weight distribution, you make a decision about which sweep or escape to attempt. Sometimes you're wrong. But hesitation is always worse than a wrong choice, because hesitation lets your opponent settle. In engineering, the equivalent is analysis paralysis. The competitor ships while you're still debating frameworks.

My rule: if a decision is reversible, make it fast and learn from the result. If it's irreversible, take exactly enough time to understand the stakes, then decide. Don't confuse the two categories.