IMHO the current state is the worst possible outcome. In addition to what we already had for most (gerrit) or some (github) of our repos, we now have a third place for some others (some tools on toolforge, Scap and a few more host their code on diffusion and do code review on differential).
Especially from a new contributors perspective, this is bad. Depending on which project I want to contribute to I now have to familiarise myself with github, gerrit and/or differential. Personally, when I started contributing here about a year ago, I didn't know about any of those and wouldn't have cared which one to learn. But by now, I had to dig into differential as well as gerrit, because I wanted to contribute to different projects over time.
This is an acceptable state for a (gerrit->differential) transition period. It isn't when the transition was declined. One can argue a lot about whether gerrit or differential fits our purpose better (which was done) - but I hardly can imagine how keeping both isn't much worse than having either of them.
tl;dr: The migration gerrit->differential being declined and gerrit being supposed to stay around, we should consider to shutdown differential and migrate existing projects from there to gerrit.
(Originally posted as T119908#4096033)
See also:
T296893: Replace Diffusion integration with Gitlab integration in Striker (toolsadmin)