A nonnegotiable feature most people want from a social network is the ability to keep certain other people out, or at least fully disconnect from all people like them - at scale, as opposed to manually blocking/ignoring specific accounts. Motivations vary, simple spam and fraud is an obvious one, prohibited material (with CSAM at the extreme end) being another obvious one, but of course at some point politics also comes into play.

You can't have a technical solution to that, it's a design choice/tradeoff. You can have a highly permissionless social network, but then it's not useful for the average person because it doesn't exclude the "undesirable" content; and you can have a highly regulated social network but then it's obviously not useful for people who do want to make the excluded content; but you can't have a single social network (even if federated) for everyone, splits are inevitable.