Depending upon your operating system of choice, there are many things that are easier to apply to single processes as opposed to multiple (killing, affinity/priority, sandboxing/permissions, etc). Also, anecdotally I've found there is overhead for each tab process (of which I may have a ton of very tiny ones in my tree tabs). One of my big use cases is that I want to embed the browser in my software. Granted Gecko is not very embeddable in its current state, but the general move towards multi-process browsers often prevents my app from being self-contained (e.g. Electron apps).
There are tradeoffs. If all of the evergreen browsers are headed this route, so be it, I just want to know.
there are many things that are easier to apply to single processes as opposed to multiple (killing, affinity/priority, sandboxing/permissions
Sandboxing is easier with more processes, at least if the processes are split up to make this easier. One of the reasons for e10s is to allow for sandboxing, so it is now easier to apply sandboxing rules to Firefox.
anecdotally I've found there is overhead for each tab process (of which I may have a ton of very tiny ones in my tree tabs
This is understood and one reason why Mozilla is conservative here.
There are tradeoffs. If all of the evergreen browsers are headed this route, so be it, I just want to know.