What non-hacker users contribute, in theory, is widespread acceptance. This comes with its own benefits, such as F/OSS solutions becoming de facto standards, which in turn means things like e.g. Linux drivers for more software, content providers not trying to lock it out etc.
However, this all can only happen if there are a lot of non-hacker users. Unless and until you actually get to that point, there are no benefits, but the support costs are there regardless.
And I have serious doubts about getting to that point, ever. I remember being much more... optimistic - or perhaps naive - back in early 00s. Many of us really thought back then that F/OSS would surely win, gradually, slowly, but surely. Now? Things are better in some ways, but overall I'd say that the environment is more hostile towards openness than it ever had been before, with widespread adoption DRM and lock-ins throughout the entire software ecosystem, massive centralization of online services, and pervasive surveillance being pretty much the accepted social norm. On top of all this, smartphones became the dominant personal computing platform, and DRM and restrictions are even worse there.
So, I don't know. Maybe you're right and this kind of "secret knowledge" for the initiated, those willing enough to dedicate significant time and effort to make it work, is the best we can have. It sure is a depressing thought, though. And I'm not so sure we can keep even this much, to be honest. E.g. at this point I fully expect to see bans on things like backdoor-less end-to-end encryption in my lifetime. Or, privacy aside, how about protocols that are straight up non-implementable because of DRM? Free software has limited usefulness if it can't meaningfully interoperate with things outside of the bubble.