You know, you're not the first person to say that about something I wrote. Neither of us use general AI, and I make a point to avoid all of those tools to do things "the old fashioned way".
It could be due to me being neurodivergent, or it could be that a certain kind of writing got slurped up by AI and that's the default style now. I don't know.
I work hard on everything I write.
So, I used messaging here in the broad sense. One possible application for it is instant messaging, which there are ActivityPub implementations out there doing that. But it can also be used for statuses or pretty much anything else that gets federated.
It's not, the demo video actually shows that being one such use case. There's nothing stopping anyone from writing a chat service in ActivityPub. But this can also apply to statuses, media, all kinds of other stuff.
Yeah, it was kind of an obscure project to begin with. Still, it struck me as an excellent way to prototype federated services that all hooked into a singular identity.
Hopefully, it inspires future work.
Nope! The main thing going for a distributed ledger would be global consistency, but there are some very negative downsides. It's slow, scales poorly, puts a massive burden on admins operating instances, and the blocks would be effectively immutable...meaning that if the contents of each activity were fully nested inside of a block, edits or deletions would be impossible. We already see problems with CSAM showing up in various blockchains, and a Fediverse blockchain used for social purposes would be no different.
It's just not the right tool for the job when it comes to social networks. There were a few early experiments in this area, such as Twister, which tried to build a Twitter clone on top of a blockchain using DHT. It has a huge list of limitations, though, and development stopped around 2020.
Silverpill might not like me making this comparison, but its form and function are shockingly similar to OnlyFans.
OpenCollective is a pretty solid open source contender in my experience. It's not perfect, but it's definitely workable.
PeerTube's support button is, at best, a lightbox that holds a text string. It's fine for highlighting links, but you still kind of have to dance around with having an account in another system to make the payment.
You might be interested in trying out Nostr. I know it's more Bitcoin Lightning focused, which admittedly is not for everybody, but they've managed to make a lot of stuff work when it comes to paying one another.
In the technical sense, Nostr is really great.
Yeah, the YouTube algorithm itself is a huge problem. I think about the fact that there are entire slop studios out there trying to ride algorithm trends, churning out crap to push onto YouTube Kids so that they'll do numbers and make a lot of money from it.
Like, I have nothing against the concept of a recommendation algorithm itself, but the relationship I just described is nightmare fuel.
Yeah, Wafrn uses both protocols, and basically runs its own PDS and AppView.
No, I just tend to break up my thoughts into different segments and structures, because otherwise the whole thing feels like I'm just rambling about a bunch of stuff.