Blaming agents for giant PRs is like blaming calculators for bad math.
Past ~400 lines, reviewers stop reviewing and start skimming. That's been true forever.
The fix isn't smaller diffs, it's better shape. One review decision per layer.
That's what stacked PRs are for:
Look, better late than never to ship stacked PRs.
But this feature was extremely useful 2015-2025 at most companies using it. Since ~Nov 2025 when AI models became more capable and write more and more of the code, stacked diffs have so much less relevance.
GH feels years behind
