One thing that sticks out is how much technical constraints defined Usenet. The decentralized store-and-forward structure was designed based on the extremely low bandwidth available (initially 300 baud modems, connected for a few minutes each night) and the various newsgroup hierarchies were largely meant to signal which groups should be prioritized over others when a site was determining what to carry. Hence the insistence on comp.women over soc.women because only the comp.* groups made it to Europe. As described in the article, the switchover from UUCP to TCP/IP made the "backbone cabal" and their carriage decisions obsolete shortly after the Great Renaming took place, yet the naming conventions remained.
A big difference between Usenet and modern services is that, except on the rare moderated newsgroups (where a moderator had to approve all messages before posting) anyone could post, all posts went to all servers carrying the newsgroup, and nobody could force other servers to delete posts. It was expected for users to "killfile" anyone whose posts they didn't want to read, but it was nearly impossible to ban someone from a newsgroup so that nobody could read their posts. Ban someone from one server and they could just sign up on another. Defederating a server, the so-called Usenet Death Penalty, was almost exclusively used against spam servers. These properties of Usenet made it an attractive platform for sharing pirated and other illegal content, which in turn led to the demise of ISP-hosted Usenet servers in the late 2000s (by which time most of the legitimate discussion groups had migrated elsewhere).
There was nothing like the current Mastodon drama with sites defederating each other over whether certain political views should be allowed by moderators, because there were no moderators and no notion that any political views were unacceptable to post. (Of course, political posts outside political newsgroups were considered rude, but there were no bans for rudeness, only killfiles...)