Is that true?
Peak IRC users appears to be 10 million (2013) - https://en.wikipedia.org/wiki/Usenet#Decline, Citation needed.
Pretending that all IRC users were in the US... That's ~3% of the population.
Usenet is arguably going strong (daily volume has exploded but isn't a useful metric here though. Can't find active users from cursory look).
~70% of US population are 'active' users of Facebook. This counts people that have only used it once in the last 30 days. And it's not a dichotomy - I don't think people have an issue stepping outside of facebook, as long as there are no barriers to entry.
Just playing devil's advocate, I don't disagree with you in general. I appreciate that my stats are lazy and blunt, but it shows how relatively minuscule the usage of IRC & Usenet were 2 decades ago, and that while direct messaging might be in a more 'walled garden' state, traffic to self-federated content is generally still as attainable. Perhaps 'eternal september' has swung so far to the other side that the old non-internet users of the 90s have funnelled through AOL and out the other side to Facebook (and notably, it's easier to grab traffic from FB users than non-internet users), while the core tech demographic has remained 'in the middle'.
Also was careful to state "stats are lazy and blunt"
But you're right that the bias has shifted, e.g. long twitter threads (with no real archivability) vs. a blog post.
Really the question of the choice is "Do you care about the audience at large or the way you and your small niche want it more?"
It is and it isn't. Sure, you can run a website, forum, blog, whatever and host it yourself much easier than in the 90s. But people look for you and connect on the popular platforms they frequent: Facebook, Twitter, Instagram, TikTok, etc.
Back in the day even normies would pick up some tech and talk to you on IRC or USENET. Nowadays, it's the opposite, the bias is in the other direction. Even hackers are on Slack and Twitter.