Been building a small SaaS for about six months now, invoice automation for freelancers, and I lean on AI coding assistants pretty heavily since I'm not a trained developer. The thing that keeps nagging at me: the code works, users are signing up, but I didn't write most of it in any traditional sense. I described problems and iterated on outputs.
That felt fine at first. Now I'm looking at scaling questions and realizing I have these gaps where I genuinely cannot explain why certain parts of the architecture are structured the way they are. The AI made a call, I accepted it because it ran, and now that decision is loadbearing.
This is probably a new version of a very old problem. People have always used tools they don't fully understand to build things that depend on those tools. But the speed at which AI lets you outrun your own comprehension feels different. The gap between what you can ship and what you actually understand closes really slowly while the product keeps moving.
[link] [comments]