I would recommend against moving code reviews to Github. I like it and use it all the time for my own projects, but for a large project like GHC, its code reviews are too basic (comments get lost in multi-round reviews), and its customisation an process enforcement is too weak; but that has all been mentioned already on the https://ghc.haskell.org/trac/ghc/wiki/WhyNotGitHub page you linked.
At least you're able to submit comments! I just had my second interaction with arc/phab and I have to say that I really hate both now. The former is not flexible enough and lacks documentation, the latter is just plain confusing. I was trying to update an existing phab diff (D1334), but I had no idea what would be submitted, and I'm still not sure whether I did that okay or not. Oh, slyfox tells me that I overwrote my previous changes, nice! In the process, I also created a new revision by mistake. The web UI didn't help either since there's so much stuff: diffs, revisions, ids. Is it okay to have multiple diffs in a single phab differential after updating? No idea. After that I was struggling to reply to rwbarton. I hit "Done" and added my comment, but both things were marked as "Unsubmitted" (or something). After a while I decided to click on the button at the bottom of the page. Looks like it did the trick, but I have no idea whether it's the right way or not. Not that I'm saying that GitHub is perfect, but at least it works instead of messing up with the work I carefully tested.