On learning by shipping

I used to collect knowledge like it was the goal. Courses, tutorials, documentation, long threads I saved and never reread. I could explain concepts I had never once put into practice, and I mistook that fluency for skill. It was not skill. It was a very convincing costume.

Everything changed the first time I shipped something real and someone actually used it. A tiny tool, nothing impressive, but it was live and a person on the other side of a screen depended on it working. Suddenly the abstract became urgent. Edge cases I would have skimmed past in a tutorial became bugs I had to understand at two in the morning.

That pressure is the whole point. You do not truly learn a thing until it can break in your hands. A finished project teaches you more than ten unfinished ones, because finishing forces you through the unglamorous middle: the error messages, the deployment that fails silently, the small design choices nobody praises but everybody feels.

So now I optimize for shipping, not for knowing. I pick something small enough to actually complete, and I get it in front of real people as fast as I honestly can. The gaps in my understanding reveal themselves immediately, and I fill them because I have no choice.

Knowledge is cheap and everywhere. The rare thing is the willingness to build something imperfect, release it, and let reality teach you the rest.