I have been bit by mailing list nonsense lately, had somebody get really angry with me for not cc-ing on everything in a series when somebody else had got angry with me for.... cc-ing on everything in a series. Sometimes you can't win.
Not certain whether a -fix patch gets picked up by bots, etc.
But on the other hand not having to deal with as you say a single point of failure is good, adding a hurdle so people have to be halfway competent to comment/contribute is good (just go look at https://github.com/torvalds/linux/pulls
for an example of what happens when you don't have that), having everything be 'discussion first' rather than clicking on a web interface is good.
Maybe if there was some git-based thing for this that was distributed and addressed issues OR of course, using tools on top of it like b4 :))
The problem with this discussion is that it's super nuanced. It's easy to get fed up with mailing list annoyances, but you can't then just ignore the drawbacks of other approaches.