Techies are just as prone to nostalgia as anyone else. In particular, they tend to be nostalgic for the time when the Internet was the domain of techies, when it was our little bubble.
We're just as prone to getting distracted by technical solutions as anyone else. In particular, when we think we've found a technical solution to a social or legal problem.
Finally, we suffer just as much from Dunning-Kruger - being so bad at something that you're incapable of seeing just how bad you are - as anyone else.
Combine these three and you get what's wrong with "The Small Internet".
1. What The Small Internet Got Right
There is a major social problem with the internet of today. It is filled with engagement bait, enragement bait, and various levels of self-destructive content peddled by the intelligence services of very unpleasant dictatorships[c]; and this is something we need to deal with.
The Internet - or a large part of it - is now not a place with educational information, but a place where you are a plaything for recommendation algorithms that only optimize for one thing: engagement. Seth Avlo describes it well in My Conflicted Relationship With Social Media[d] - even if you set out to consume content to better yourself, you are exposed to a very powerful system, created by people smarter than you and proven to work, whose goal is to turn you into a doomscrolling zombie.
Wishing for the simpler Internet, where you could find good stuff without being mentally hijacked, is a very understandable wish. By and large, the arenas of the commercial Internet where people (average people and professionals) can publish and contribute to the world - various social media platforms - are driven by clicks, likes and engagement, if not directly on the platform, then by platform engagement driving traffic to their website.
2. What's Wrong
But if you want the Internet to return to the "cozy community" stage that it may or may not have had at its infancy then your thought should not be to create new protocols that enforce your idea of a cozy community. If you're a techie and the only hammer you have is the ability to write code then this may sound good, but like many technological solutions to non-technological problems it fails to actually solve anything and instead messes up a lot of other things.
Just how getting rid of images and tables would result in a community was never really explained, but I think it's a logic error where someone confused the style of the infant web and substance of it.
What Gemini did was create a cozy community around the protocol, but like other communities created around a specific technological solution, once the solution either hits a dead end or becomes too complicated for the average member to understand there's nothing left. And it does get complicated, but when you're an enthusiast just starting something and all problems are easy, you can Dunning-Kruger[e] yourself into thinking it won't. (HTTP, the protocol used by the web that you are browsing now, is the result of over a generation of research by people who are very much not Dunning-Krugered.)
3. Is There A Solution?
I don't have an answer, except to stay away from the engagement-driven Internet and build the community you want using the existing software stack. Running your own server is something only a very miniscule minority will ever do (I'm a professional and I'd never do it), although I obviously do like having this blog of my own.
But staying away from the engagement-driven Internet is something you can do without inventing new protocols or throwing away a generation's worth of progress.