I get where you're coming from, and I have largely the same sentiment, but it's obvious that you've never had any experience running a platform, or using one with no/few rules. You need baseline moderation, even if it isn't super ideological. You've got to keep pedophiles and misanthropes out at the very least.
Sounds like I really should be building one with the AI moderation I have in mind. I kind of have been to random deep ends. Most of them can't be that hard to filter out or steer them away from undesired content, especially in this age of AI...
Because, if it's purely about filtering out content not desired by users, it could be nearly trivially done at the edge, automatic and completely de-humanized, and the word as appearing lately doesn't read that way to me.