These people are an opportunity for the rest of us to be better people, to show compassion and mercy. They are not worthless to the rest of society. Giving them LLM companions says that they are — that their only self-worth is their feelings.
The thing is, individual moderation isn't sufficient to combat collective capture. This is Ivan Illich's idea of "radical monopoly", where a technology (like cars or the internet) becomes so entrenched that individuals can't realistically opt out any longer. This happens when we abdicate our collective responsibility to the internal logic and incentives of the tech.
I don't know the answer. But a robotic voice is probably not a bad idea — just having a reminder that the thing you're talking to is not actually anything like you. If you want to go full send, you could have the LLM generate a clickable interface on demand so you could interact with it as a machine. Voice/computer interfaces are obviously useful, especially for disabled folks. But the ones that existed in the past didn't pretend to laugh at your non jokes or imitate vocal fry.
Right, I think the challenge is that LLMs are essentially language-based, which is itself a very convenient interface. Covering the maximally humanistic default interface with something more mechanical is like tying your own shoelaces together, but it would protect us from the psychological hijacking we're so prone to when interacting with these machines.
This is the opposite direction AI should be going. Human relationships are the most valuable thing we have, and so, naturally, technology seeks to intermediate and now replace them.
Europe is adopting open source and open protocols, not to promote individual sovereignty, but explicitly to protect European sovereignty from foreign influence. This is not what these technologies were built for; "promoting democracy" does not protect the rights of individuals.
The technology listed is mostly federated, not radically open (like, for example, nostr). In particular, ATProto has provided the EU with the perfect opportunity to signal openness while simultaneously standing up a new walled garden in which dystopian "moderation policies" will be the norm.
I have been telling people lately that I feel like I'm losing my mind. And I'm not even someone who has leaned into AI coding that much either; I've just tried to learn the tools since Claude got "good". But my inherent laziness, which was always flattered as something that makes me a good programmer, has made me unable to use the tools with the required discipline. The result is that I have not thought deeply about the software I write for around 3 months. Every additional week that goes by without me doing a refactor or serious feature addition saps my confidence. I know I can still code. But I feel worried that I can't. Today I am refactoring a 4k LOC AI-written rust codebase. I don't know rust, but I will finally learn it today. And I can already tell the end result will be 50% the size and immeasurably more coherent.
The basic idea is that you can put your repository on multiple GRASP-compatible nostr relays (GRASP is a sub-protocol that glues nostr and git together), so even if one server goes down you can transparently sync using the others. This means in effect 100% uptime if you choose reliable servers, as well as cryptographically-signed repositories, activity, issues, etc.
Yeah, any time you need either an index or a caching layer you have to re-centralize one way or another. But decoupling those "services" from the data storage itself helps, and credible exit makes the gatekeepers far less powerful. An example: a few weeks ago nostr.band, one of nostr's main indexers/search services went away. Search is still somewhat impacted (evidence that we were centralized around it), but indexing (i.e. finding users' relay lists) is still covered by several other services.
Getting clients to do the right thing is like herding cats, but there has been some progress. Early 2023 Mike Dilger came up with the "gossip model" (renamed "outbox model" for obvious reasons). Here's my write-up: https://habla.news/hodlbod/8YjqXm4SKY-TauwjOfLXS
The basic idea is that for microblogging use cases users advertise which relays their content is stored on, which clients follow (this implies that there are less-decentralized indexes that hold these pointers, but it does help distribute content to aligned relays instead of blast content everywhere).
Also, relays aside, one key difference vs ActivityPub is that no third party owns your identity, which means you can move from one relay to another freely, which is not true on Mastodon.
Is it open source? This is a cool idea, but I'm pretty sure it's probably just a thin wrapper around claude. I also couldn't install it on my headless dev box because it relies on a localhost callback. Well, I'm looking forward to the first open source version in about 10 minutes.