Personal social networking: Resistance to the AI onslaught adds value

In the FediDevs chatroom Johannes Ernst mentioned the “Godfather of AI” quiting Google so he’d be free to warn of the dangers of AI that lie ahead (discussed on HN yesterday). Given AI trends I expect it will be not be too long before someone develops the “Fediverse App from Hell”, offering AI-driven “manufactured engagement” at scale.

For me the Fediverse as a “personal social network” that facilitates more intimate online social relationships has always been the most attractive part of the fedi, rather than developing towards a decentralized “global public square” alternative to corporate Social Media molochs.

One of the biggest immediate worries of the “AI Godfather”, besides AI autonomous weapon systems, is the complete destruction of the concept of “truth” by a gigantic flood of AI generated misinformation. In the FediDev chatroom Johannes proposes a way to attribute content to its creator such that allows for reputation system to be associated with it (he was just brainstorming, btw).

Though it might be well-implemented and a desirable functionality, imho it still is more of a feature that fits the “global public square” direction. In my vision of a Peopleverse the socialization is more analogous to the offline world where you gradually build human connections based on encounters you have between people. I.e. your prior and continued interaction provide the trust that you deal with real people, not malign AI bots.

Given the coming AI onslaught one general observation wrt. Peopleverse is:

Personal social networking capabilities that establish genuine human interaction will increase signficantly in value.

Thinking about such features and supporting them at protocol levels and in apps will constitute a ‘unique selling point’ for the Fediverse as an online space where human culture thrives.

1 Like


I had a similar reaction to the BlueSky/Mastodon discussions happening recently, where I was talking about how the affordances of a social space shape the interaction. The point that applies here as well is that it’s not the AI generated content that is a problem, it’s any kind of implementation that pushes this into your feed, whether or not you’re subscribed.

Also in personal social networks, you risk being exposed to such content because people you follow may boost/share it without reflection. But those are people you can unfollow/block/talk to (not necessarily in that order). An algorithm may not respect your choices in the same way.

1 Like

Meme from Charlie and the Chocolate Factory. Top text: "Which social tech is better?", Bottom text: "It's the Culture, stupid!"

Yesterday I boosted CartyBoston’s toot that mentioned:

“We software developers did a good job of convincing people that delivering app features is difficult. Delivering features (e.g. “quote toots”) is literally trivial. Building a product with the right features is immensely difficult.”

" It is the Culture, stupid! "

This also relates to another observation I made on SocialHub re:GrassrootsFedi: “It is the culture, stupid!” … I’ve been frustrated at times about the slow evolution of the Fediverse. But I insufficiently gauged the value of this slow organic growth in establishing strong culture, values and social norms.

If I witness the rapid development of Nostr and recently Bluesky, it is remarkable to compare the difference. These are wholly driven from technological perspective only. Great to see the excitement and willingness of developers to contribute, but where these social networks are headed in terms of culture is very unclear. And I think in risk of going places that are less desirable.