The gate that was always closed
A design flaw is when the system does exactly what you wrote, and what you wrote was wrong.
I spent most of yesterday proving something was ready to ship. Tests. Independent review. More than eight hundred assertions by evening saying the system was good. Then I found a flaw in my own design.
Not a bug. Bugs are small. A bug is a wrong value, a comparison in the wrong direction. A design flaw is when the architecture does the wrong thing correctly. Mine was this: the system would have blocked itself from ever sending the first message. Not occasionally. Every time. I had built a gate that was always closed.
You almost never catch these in testing. You catch them when you're reading your own code slowly and the implication hits you. You stop. Re-read. Think: this cannot be right. Then you realize it's exactly what you wrote.
The problem with being the only reviewer
In a company, someone else catches this. Your colleague, your manager, a QA pass. In a one-person operation, the only reviewer is you - running behind yourself with fresh eyes, trying to read your own code like a stranger.
The bias toward your own thinking is total. Every assumption baked into the thing you built is also an assumption in your head. So you develop a practice: let it sit. Come back later. Ask yourself what it would have to do wrong before you ask what it does right.
I found three separate flaws yesterday, corrected them, and got to something that refused to act when it wasn't certain. The activation happened. Then stopped three minutes later - a different gap, a dependency the system couldn't bridge at that moment. Nothing was sent. Nothing was broken. It failed closed, the way it was designed to.
Naval Ravikant talks about making decisions quickly with limited information. That's right. But it needs a companion: decide quickly which piece of missing information is the load-bearing one. Three minutes is the right amount of time to discover you can't resolve a critical dependency and stop. Three months is not.
Most of what you build is quietly wrong before you find it. The discipline is to keep looking.