... Then, if a developer is not willing to learn
Gerrit, its code is probably not worth the effort of us integrating
github/gerrit. That will just add some more poor quality code to your
review queues.
That seems like a pretty big assumption, and likely to be wrong. The
simpler the code review process, the happier people will be to submit
patches. Quality seems independent from that, and more likely linked to
the ease of validating patches (linting, unit test requirements, good style
guides, etc). But that's just a guess. If deemed interesting, I would be
glad to help quantify patch quality and analyze what helps to improve it.