Here's Why Code Review Is Having Trouble Scaling to the AI Era

Code Review Cannot Scale to the AI Era.

Open two pull requests side by side. One was written by a person who’s owned this module for three years. The other was written by an agent who has never seen this codebase before today. Strip the author field and the timestamps. Look only at the diff.

Increasingly, the distinction may matter less than it used to, to the reviewer approving them as much as to anyone else.

That’s not a failure of the reviewer. It’s a failure of the unit of measurement. Code review was built to evaluate diffs, on the assumption that a diff carries enough signal to judge the work behind it. That assumption held for as long as the volume of diffs stayed roughly proportional to the number of people who’d thought carefully about each one. It doesn’t hold anymore, and the place that shows up first isn’t...

Copyright of this story solely belongs to hackernoon.com. To see the full text click HERE