I once worked in an org that shipped massive pull requests as the default. Not occasionally...as a matter of habit. A single PR could sit open for days (or even weeks in some cases), because reviewing it meant holding an entire subsystem in your head at once. Bugs piled up. Deadlines slipped, again and again. And it ended the way these things tend to: we eventually had to rebuild a big chunk of the system, because it had gotten into a state nobody could safely change anymore.
Here's the part nobody said out loud. The engineers weren't bad. They were smart people working hard. The PRs got big for a much more boring reason...and it's probably the same reason yours do.
Nobody ever taught them how to cut the work down.
Big PRs are a skill gap, not a character flaw
We talk about large PRs like they're a discipline problem. "Just make smaller PRs." As if the only thing standing between a 1,500-change diff and a 150-change one is willpower.
