Building a thing and proving it works are two different days

There’s a quiet dishonesty that creeps into building, and I’ve caught myself in it more than once. You finish something, you ship it, and you let the good feeling of shipping stand in for proof that it actually works. Those are two different things, and the gap between them is where a lot of products quietly fail.

So the system keeps them apart on purpose. Cutting a version means one thing only: the work is built and in people’s hands. Whether it works is a separate question with a separate bar, and that bar is only cleared by real evidence from someone who didn’t build it. Someone outside sets it up on their own and reaches the point of using it for real. Until that happens, the version is shipped, not proven, and I make myself say it in exactly those words.

The two don’t happen on the same day, and they aren’t supposed to. An early version was cut on one day and didn’t clear its bar until several days later, when the proof finally arrived from outside. That waiting period is uncomfortable. The thing is out, it looks finished, and everything in you wants to call it done. Calling it done is the mistake.

I built this rule because I don’t trust my own optimism. Left alone, I round up. A thing that mostly works becomes “it works.” A promising sign becomes proof. Twice I let myself declare something proven on evidence I didn’t actually have yet, and both times I had to walk it back. The rule exists to catch the version of me that wants to believe the flattering read.

The idea I’m rejecting is that shipping is the finish line. Shipping is the start of the only test that matters, which is whether the thing survives contact with a real person who isn’t you. That test has its own day, and it comes later, and I don’t get to skip ahead to it.

This is also a promise about the numbers you’ll ever see me put on this product. If I tell you something works, it means someone who isn’t me proved it, not that I shipped it and hoped.