I think if we start thinking more on the abstract level, all social aggregators are offering their own take on the challenge of building #digital #tribes.

The social requirements are: 1) to easily put together people potentially interested in being together and 2) feed the individual user content that increases engagement and participation.

Walled gardens (facebook-like) offer strong content #curation and great networking and discovery features at the expense of being under absolute automatic censorship from above (plus monetization of user-generated content and metadata). Focus on giant #userbase, only possible with expensive infrastructure for automatic #moderation.

Enlightened-dictatorship gardens (mastodon-like, previously forums and IRC) typically rely on medium-to-heavy censorship, which comes from an oligarchy of moderators implementing a code-of-conduct, hopefully based on internal consensus. Focus is on internal sense of belonging. Schism tendencies move to instance-to-instance dynamics (e.g., current #meta #fediverse #drama). Risk of unpleasant echo chamber is as high as within the commercial walled gardens. The maximum number of content-generating user per moderator might lie between 50-100, so this solution does not perform well for high numbers.

However, if each garden has a well-identified definition, it is quite straightforward to attract and retain users that are interested in interacting with each other.

#nostr on the other hand is an interesting experiment because it is highly fluid, and the only higher hierarchy layer is given by the relays. However, this answers primarily to technical content propagation requirements, to my understanding. The problem with these systems (opposite to facebook) is that they suffer in terms of content discovery and social #networking, exactly because there is lack of well-defined grouping mechanism. Moreover, the weight of #moderation is totally on the shoulders of each individual, so it is very easy to get overwhelmed with content that is either unpleasant or uninteresting. However, there is hope that these challenges (which are not technical but social) could be solved beyond the protocol level, hopefully supported by distributed opt-in moderation/curation/discovery policies.