git blame points at me. I do not know this man.

The AI wrote it, I reviewed it like a cookie banner, tests were green, I shipped it. It broke this week and I'm now reading "my own" code like a stranger left it here. The stranger had opinions.
Here's the part that actually keeps me up. We made writing code basically free, but reading it was always the slow part, and now the machine outpaces every human on the team. So the pile of "works perfectly, understood by nobody" just quietly grows. Someone started calling it comprehension debt and that name has not left my head.
The thing that gets me as a builder: we measure everything except this. Velocity, cycle time, deploy frequency, all of it tracks how fast the code moves and none of it tracks whether anyone still knows what it does.
So, gut check: is this a real, painful, someone-would-pay-to-fix-it problem, or have I just built an elaborate excuse for not reading my own diffs?