I've been finding the increase in PR review volume extremely exhausting, partly because it's just way more code than typical, partly because it's just extremely mind-numbing work, but also because it's emphasizing how it wasn't necessarily the physical act of typing the (most) code that was expensive mentally, but rather that it was internalizing the parts of the system to a sufficient degree that I know why the code should have changed. With 5 different (large) PRs in 5 different subsystems, potentially each day, I'm now re-hydrating my own mental context more often, but not in a way that actually helps me become sufficiently intimately familiar with those systems, because that did come manually writing the code. My connection with the logic feels more superficial than ever, like I'm kicking more and more details out of my brain simply because of the volume of either more shallow tasks, or one-off specific tasks in an area I haven't seen that I don't end up sitting with for long enough
Hacker News