As a webmail service it's quite usable; on the front-end it's well designed and the performance is getting better. While to my knowledge Tutanota is the only comparable service to have as of yet deployed mobile apps (iOS, Android), ProtonMail has theirs upcoming as well, also their recent certificates upgrade seems a good sign. From a branding and PR perspective they've had a very good start, rallying buzz and an enthusiastic following. For the time being I'm enjoying following it and testing it out against others e.g. LavaBoom when it comes out more etc.
That all said, yes, all of these new players are still building JS-based encryption on top of email as it fundamentally exists today; no metadata protection... and at this early stage, for them taking market share is kind of like battling for the middle-ground between how much convenience people will be willing to give up, more to get out of Google's ecosystem perhaps than anything else, without asking them to change their email behaviors fundamentally... however large or small that total addressable market might be. I for one find myself applauding more the spirit of them than using them in practice, because that leads me to want to have more convenience and functionality out of them than they can currently offer (e.g. "Oh, but I can't connect my mobile mail client to it"... "Oh, but there's no integrated calendar that I can moreover read / write to from my mobile devices"... "Oh, but the 2 deeply email-integrated CRMs I use (each for a different requirements set) can't connect to it" etc.). So it all ends up feeling like an increasingly usable, increasingly attractive-looking and good-feeling, class of tools that nonetheless don't fundamentally give the desired level of privacy and security options that brought them on my radar to begin with. I end up just reminded that for many things, it's kind of hard to get folks out of Google's ecosystem if they've been thoroughly pulled into its many services for a few years, partly per many 3rd parties having built their things off of them. It does, however, prompt one to think critically about how much one really needs and wants secure email, similar to how SilentCircle's suite of tools do regarding use of encrypted phone calls or text messages, and in so doing to make decisions about what constitutes security or not.
One interesting read:http://paginas.fe.up.pt/~ei09128/2014/07/protonmail/
In any case, and certainly the more surveillance stays in the news, at least some of these platforms likely keep growing when many consumers want to have their cake and eat it too between privacy and convenience (and it's a question of that more than one of usability), similar to how they want to between privacy and (free) publishing / sharing capability on social platforms. Email has conditioned people out of the patience for the "mail" part i.e. something one can handle waiting until one gets home or to the office to check, and arguably out of developing good reading and writing skills a bit too.
On that note perhaps one fringe benefit, or side effect, of DarkMail will be the attraction of a class of users whose levels of average literacy hearken back to the days when mail used to be the actual writing of letters... The irony of course being that there would be no way to evidence for or against that either way.