[mattcameron]

This is 3.2.2 in a small series of notes on launching an app built with AI.

AI makes it easy to generate code and request reviews. But if the reviewer has to guess what you decided and why, then the feedback drifts into opinion instead of being useful.

Simple test:

Can you name one decision, one trade-off it created and one question you want the reviewer to answer?

If not, the review is not ready to request yet.

For me, if I cannot explain a decision easily, it usually means I have not thought it through enough yet. That is the test really. Try and explain it. If it comes out clearly, then ready. If not, there is probably more work to be done before asking anyone else to look at it.