It's funny. At Dropbox, engineers *have* to merge their own work (we call it "landing"). When I first got to use GitHub (for asyncio, IIRC) I could land other people's patches once I approved of them. That was very satisfying, and felt like the "better" workflow, and we adopted this for mypy. Cases where it's better to wait for the author also exist, but it's always clear from the author's comments. In general merging sooner means fewer conflicts with other work, and we like that.

But I'll refrain from merging CPython patches by other core devs since it seems the consensus that usually the author (if a core dev) wants to merge their own work on their own time.

--
--Guido van Rossum (python.org/~guido)