Response to "Go and build it" - Why the sorv algorithm is best implemented on a pre-existing site

As a passionate advocate for random-sample-voting to build meritocracy, I am frequently asked why I don't just "go ahead and build" a site that implements it. I'm focusing on trying to get existing social media sites to implement the algorithm (or, in the long run, getting the next generation of social media sites to incorporate it), for several reasons:

Mostly, I think "go and build it" misunderstands the nature of the problem - I think the lack of meritocracy on social media is a security issue that should be fixed. If someone discovered a publicly exposed non-password-protected database on Instagram's server, which allowed users to hack their own post's "likes" score and boost visibility, nobody would say "Sounds great, why don't you go build an Instagram competitor that fixes that bug" -- people would understand that Instagram should fix the problem. The non-meritocratic nature of social media allows people to do basically the same thing (either by gaming the system or getting lucky), which is also an argument for fixing it.
Two of the most important benefits of the sorv algorithm -- non-gameability (it's difficult for bad actors to manipulate results by signing up with bot accounts) and average-high-quality (the user sees high quality content on average, because "bad" content only wastes the time of the initial sample, while "good" content gets pushed out to the entire target audience) do not kick in unless the user base is large. Thus, a proof-of-concept with a small number of users would not in fact prove anything.
But, suppose I (or someone) did go ahead and implement sorv on a new social media site. The real problem is that creating a new social media platform and attracting lots of users requires, bluntly, a lot of luck, regardless of the objective "merits" of the system that you're building. (A typical social media site costs much less to build than it rakes in in profit -- which is how the founders become billionaires -- and if it were possible to do this without a lot of luck, then Silicon Valley would be launching a new successful social media site every week, instead of once every several years.) The simplest explanation is that most users want to go where all the other users already are (the "network effect"), which makes it very difficult for new competitors in the space to attract new users. Facebook displacing MySpace was a once-in-a-generation upheaval.

Reddit, in fact, gives away the source code of their website for free; in an extremely naive view of "meritocratic competition", you could just set up a copy of Reddit and instantly have just as many users! Nobody believes that, of course, but more interesting is that nothing even approximately like that has happened -- there is no Reddit clone that anyone has ever heard of that was set up using their source code.
If a brand-new social media site based on sorv fails, it is likely to be viewed as a referendum on the merits of the sorv algorithm, rather than due to the fact that you have to get really lucky to launch a new successful social media site.

For these reasons, I'm focusing on: (1) trying to get an existing social media site to adopt some version of this; and (2) injecting the idea into the zeitgeist generally, which increases the chances of it being implemented in future social media projects.