buc.ci is a Fediverse instance that uses the ActivityPub protocol. In other words, users at this host can communicate with people that use software like Mastodon, Pleroma, Friendica, etc. all around the world.
This server runs the snac software and there is no automatic sign-up process.
Personally I'd hesitate to use LLMs in the chatbotty/generative-y ways they tend to be used even in the idealized case you outlined. I agree you can probably eliminate most if not all of the myriad objections raised against say OpenAI and ChatGPT. What I think you can't easily eliminate is the de-skilling, psychological degradation, and other such harms that offloading thought to a machine will bring. Overuse of generative AI is a kind of sedentary lifestyle for the mind.
The analogy with rabies is also wrong, but in that wrong-but-so-right kind of way.
My thesis here is that this is an unstoppable coalition. Which is good news! For the first time in decades, victory is in our grasp.I don't think the "victory" of this so-called "unstoppable" coalition would be good news at all. What he seems to be describing is taking the bougie white internet global. Laissez-faire justice that won't trickle down to anybody.
I feel somewhat sorry for people who are deliberately letting themselves be left behind by ignoring a technology that fundamentally reshapes how cognitive tasks are done in the workforce.This sentence is a seven-layer-burrito of wrong. It's quite amazing how nearly every clause is divorced from reality, and then wrapped in a tortilla of condescension.
I suspect people who lean on LLMs too hard for too long start to lose their coding/syntax and writing/generating skills, while potentially continuing to reinforce their computational thinking and reading/recognizing skills. That might explain some of the observations that LLM use can make people less productive even as they believe they're being more productive.
Of course, perhaps if AI can do all the things its boosters claim and corporate America buys in, maybe it doesn’t matter what anyone thinks? It will generate enough money to keep the politicians happy and winning elections. This vision of what we might call escape velocity is baked into almost all AI boosterism: the escape from the mundane stupidities of public accountability or public oversight.To put this differently: "if AI can do all the things its boosters claim, the elite strata of (American) society can decouple from the public and start doing whatever they please, accountable to only one another". To me this sounds like a recipe for generating a mass revolt.
The fediverse software I run (snac) puts the muted hashtags and the control thereof front and center. I fairly fluidly change these as the days go by and I change my mind about what I do and don't want to see. I like this feature quite a bit.
I can only imagine how much the engineers working on this must hate it, too, but I imagine changing it at this point would be an absolutely Herculean refactoring and migration effort. This is exactly the sort of tech debt that just gets kicked down the road forever..My God yes, I can see it. I can hear the meetings about it in my head. I can see the hundreds of 5-year-old Jira tickets. I bet they have internal Slack channels just for griping about it!
Oof that really does sound like a PITA. I don't understand why they didn't have workspaces be something you're authorized to work with once you authenticate. It seems like a flow that would eliminate so much misery. You can still do magic links etc.