#programming

When an agent helped write a pull request, the first question most reviewers ask is who wrote it. I think the better first question is what it touches. Here's how I'd set review depth by risk, where authorship still fits, and why some changes may need very little review at all.

When your team stops trusting its AI code review, the obvious fix is to add a second AI. I don't think that buys you much...two models trained on the same internet share the same blind spots. What actually helped me was checking the evidence outside the diff. I tried it on a real queue of 58 open PRs, and here's what it caught.

On a lot of teams, code that a pair of devs wrote usually gets a lighter code review. Nobody wrote that rule down and nobody voted on it, but pairing earned it. Here's what the research actually says about that discount, why it's never one flat size, and what happens when dev + agent starts getting filed under the same category.

When making something gets cheap but judging it stays expensive, the cost slides downstream...social media did it to your feed, AI code generation does it to your repo. The difference is a bad post washes away by morning, and a bad merge sits there for years.

The risky PR is rarely the big one...it's the one carrying money code, a migration, or a red build, regardless of size. Here's what I look for when one lands, and the product I'm building so senior attention goes where the risk actually is.

Deleting a year-dead feature flag turned a forgotten feature back on and broke production...here's why flag cleanup is a runtime change, not tidying, and how the context you need decays while the flag sits there looking harmless.