The AI code-review bottleneck — 200× more output than reviewers
AI generates ten thousand decisions an hour; humans review fifty. Where the throughput mismatch hits an enterprise and what has to change in the review process.
The trade-off I'd flag
Review capacity, not generation speed, is now the constraint. Teams that do not redesign review end up with faster-written code that ships slower.
