Really hoping that proliferation of instances will lead, in time, to divergence in modeling (at the software level) the parameters of "community" and sociality.
GNU Social/Mastodon shares features with SMTP, IRC, Usenet (surprised not to have seen that mentioned by others), with BBS networks before that, and even such antiquated systems as postal newspaper exchanges. In this history, "social media" giants are distinct only in their resort to centralization.
The organizing model is unchanged.
What GNU Social/Mastodon may afford is a decoupling of network application from network effect. In particular, Mastodon builds on the network effects set down by GNU Social, in the same way that gmail and MS Outlook build on the network effects of SMTP.
This means that the network application instances can diverge, in how they model and render the underlying network, how they integrate and leverage tools and tooling, so to present different use cases for participants.
My fervent hope is that some of those divergences fundamentally question the model of network "spaces", this itself a derivative of the commonly held notion of a "public sphere" (per my reference above to postal newspaper exchanges).
My suggestion here is that the episteme, the condition of possibility, if you will, that informs our discourse on what and how so called "social networks" can enact relationships between human operators of computing devices is deeply anchored in an earlier epoch.
"You do you" is Eu-gen-ics.