@fyrfli Interesting. I think I've seen some services for spinning up mastodon and pixelfed instances with fair fees. I forget the names though.
Is there something specific about gotosocial that you prefer?
I've been working on it from a different side with statically published blogs being available on the fedi via @sutty and distributed.press.
We still need an "app" for folks to easily load followed AP data into their device instead of large instance databases.
@mauve the modern conception of "self hosting" generally involves paying AWS for all the hard lifting, obscures the actual infrastructure, and hides the cost of maintenance. true self-hosting in a way that empowers us, rather than subjugating us to a megacorp, entails a degree of understanding of the technical AND the social aspects of infrastructure and maintenance.
@atax1a Honestly I hope we can get to a point where you don't need to think as much about infra to self host. Kinda the reason I've been looking at using #p2p publishing where you wouldn't need to set up dns or mess with ip addresses and be able to run everything you need right from the devices you use.
@mauve @atax1a Exactly this. I want protocols so damn decentralized that my *phone* is a fully functional node.
I would accept a transitional phase where I buy a cheap network appliance (like $100 level) that stays at home, connected to the internet, as my “personal server”
But all this multiple-container, monolithic web app architecture hasn’t reached that stage yet.
@mauve @atax1a And, since colonizers are sneaky fuckers, I should add:
If there is a transitional phase, involving my having a cheap personal network appliance, that appliance DAMN WELL BETTER BE SELF CONTAINED
(Variations on this theme, observed during the cryptocurrency fad, often meant your “personal server” was just an edge node for their cloud. WRONG!)
@fleeky @cmdrmoto @atax1a Yeah, I still love SSB. The approach to social graphs and moderation is still my fave of any social networks I've used. Also of course love the community and the pals I made there.
One thing I don't like is the data model. Needing to fully replicate logs and build up local indexes is bad UX and I think sparse pre-indexed sparsely replicated datasets would improve things massively. (also the use of json and weird node dependency management was pretty rough).