The content for billboards is also "generated by a relatively small number of users", but those users still have to pay to use them.
There’s certainly some marketing (“Check out our new paper”, “My lab is hiring”, etc), but it’s not particularly top-down; everybody does it to some extent. It’s also mixed in with actual scientific discussion, lab “hacks”, and some silly banter/blowing off steam. All in all, it’s been pretty helpful for me careerwise. I think I’ve gotten more useful advice about grant writing and papers to read than many more formal mentorship arrangements.
There are a few “blue checks”, but they’re mostly a) early adopters or b) people with some kind of public presence. As far as I can tell, neither I nor most others care about the checkmark per se. I talked about my research (twice!) with John Carmack and it was cool to know that it was really the Doom Guy, but I think I probably would have engaged with anybody equally curious about my work.
Because the content on Twitter is generated by a relatively small number of users. A lot never tweet, quote tweet or retweet but it goes beyond that number once you weight it by audience. A small number of people have a large amount of reach and thus are responsible for a good chunk of the content.
That content is why the users are on Twitter and it is those users who are advertised to that pay for Twitter to exist. That's what Stephen King means.
Now you can turn this around and say that those power users are only there because the audience is and that they get value for being there but platforms need users and users follow creators more than the platform as a general rule.