worse still — we aren’t even talking about socialization. What Sanders is talking about is (partial) nationalization. The question we are debating is not “Should the workers take over?” but “Should the government?” It may have escaped the notice of those in that debate, but the current government is not particularly sympathetic to socialist interests. That Donald Trump and Bernie Sanders have both “called for the nationalization of AI firms” is presented by Guastella as a positive thing, as transcending the partisan divide. We think it represents a grave strategic error.
This is particularly crucial precisely because AI is attacking many of the most privileged workers, bringing the logic of mass production to realms that have so far been relatively protected from it. Take, for example, therapists. It’s easy to picture the capitalist AI future for mass-produced therapy: it’s not the rosy picture of full automation, in which a chatbot can provide competent therapy for the masses. The future in which we all have dozens of “agents” working on our behalf (or whatever the current sales pitch may be) will look bleak — like a single therapist overseeing window after window of AI chat interfaces. Client after client is talking to a language model, the trained professional is there to oversee the chats — scanning them for dangers and possible liabilities. The provided care will be of dubious quality — but boy, will it be cheap.
(still guild-related thought pattern (is it proudhonism or whatever it’s called?), but at least approaching humane thought rather than previous drivel (linked at the top of the article)
no mentions of ai weapons systems, of course.
I’m going to scream at luddites
You can restrict usage of things, you can limit use to efficient models that won’t destroy the planet, etc., it’s not a choice between “use it and let it run and ruin everything” or “destroy the lying machines.” You don’t need to use it for fucking therapy to get benefit out of, say, food workers not having to man intercoms to take orders. There are labor and time saving utilities here but i swear to god half of lemmy hasn’t interacted with an llm since 2023. If we’re talking about a hypothetical where we’re nationalizing this shit we’re talking about a situation in which we’re past being in capitalist hell where every implication of llm use is through the lens of bourgeois control, have a little more imagination than “we must smash machine”
I am sure it can do some impressive things, though I can’t say I have seen great evidence of it (where it can do impressive things are probably fields I don’t understand). I also think it’s weird to call people luddites over this. Like yea, if we have communism and it’s used for good, sure why not. As it stands now though, to me it reads like “How we can use the torture nexus for good assuming it doesn’t kill all of us” which doesn’t inspire me.
You’ve got the west essentially trying to make a turbo fission powered super calculator because they think if they drain the biosphere to power a computer it’ll make it intelligent. China is not doing that with their research and they’re developing models nearly “as good” that work with a fraction of the resources. And that’s talking about LLMs doing shit like writing code, which isn’t what I think LLMs should be used for and not what I’m saying when I say “they have uses and utility.”
It’s probably the most apt comparison that could ever be made because of people wanting to smash the machine instead of seize and utilize it
Also I’m not necessarily saying the results of current LLMs are “impressive,” like, don’t get it in your head I think Dr LLM should be a thing since I literally just said they shouldn’t be used for stuff like therapy.
But like in my own case I had terrible neck and shoulder pain progressively getting worse. I was (and still somewhat am) worried it was related to my psoriasis, like, oh great, degenerative spinal condition time. The doctor I saw basically had no time for me and just gave me a referral for physical therapy and a muscle relaxer that she and the office pharmacist gave 0 explanation for.
A few random questions on google search later (i honestly don’t remember how i formulated the query) and the LLM is basically like, is your mattress memory foam? Is it on perfectly rigid support? And it turns out, no, it’s on a box spring because my dad didn’t want to give us the support it came with. The mattress was sagging in random points between the springs and was literally destroying my spine.
I put it on the floor and instantly felt better and now 2-3 weeks later my neck and shoulder feel normal
Then the muscle relaxer turned out to affect how my CNS deals with adrenaline. Which I found out not from the doctor or the office pharmacist but as I was actively undergoing withdrawal symptoms. Fun! The best part about that was complaining about it on here to have Nerd Scolds tell me not to use LLMs for medical shit.
Let me say this as a separate sentence for those nerds: i asked the doctor and the office pharmacist for info about how to take it and was given no caution. The doctor was less informative than the llm. Idk how people can read that and still tell me “don’t google shit” like the alternative (knowing nothing due to the human doctor’s negligence) would be better.
Then I’m typing in shit about it being humid in my house despite the A/C and a dehumidifier running and it starts asking me about the AC condensation line. I still don’t know if it’s the cause of the humidity, but it pulled up the instruction manual for my AC unit which clearly indicates the P trap on the condensation line should be deeper, it currently has a 1" water seal but the manual calls for 3". And it’s definitely sucking in outdoor air through the p trap, so i think it’s right in that it’s a problem even if it’s not THE cause of the humidity issue.
Those are 2 things where I might have stumbled on the necessary information through reddit threads on a regular google search, but tbh I doubt it because I wouldn’t have put in the time to continue interrogating those issues. And then the muscle relaxer thing is an instance where a human professional (two, actually!) completely failed to provide me critical information on a drug’s safety.
All of that is shit i would consider “useful” and that’s why i’ll always get irritated at people who act like LLMs are stuck in 2023 giving you recipes for Glue Pizza.
I don’t think any of these uses are worth the extremely wasteful way the West is doing this shit, though. But that doesn’t matter in entirely hypothetical situations where capitalism is abolished and we’re nationalizing shit
Yea honestly the situations you are describing are some of the more common use cases that can be actually helpful. I do wonder if google and the others getting so fucking shitty amplifies the feeling. But I have heard the same from people I trust. Like you said, can be good stuff (the computer from Star Trek basically) but have to be able to do that without murdering the environment even more.
Oh no definitely, like, I wouldn’t say it’s better than the search results themselves were a decade ago before they fucked it all up. But it’s still presently a general improvement over “search + reddit” to find even a scrap of info
A data center is being put in near where I live and they have flooded the houses there.
Then we have flock cameras coming to my town.
Like I don’t think you are wrong. But you have to meet people where they are at. Its hard to see the good in something when the only thing it is being used really in America is trash images and shit and making things worse.
People are afraid that if they let anything through it will lead to the parts that is causing actual harm. And they have every reason to feel that way given the mountains of evidence that shows you can’t have anything good in this hell country.
The people are responding the only way they can think of to such a thing. Its our job to show another way.
Also this.
I do believe Ha-Joon Chang’s spiel about how the washing machine has been more transformative than the Internet, but when I went to live with quasi-Luddites, they were doing pretty fine with their half-barrel hand-cranked washing machines.
Before 4 years ago, “AI” meant programming, it meant a kind of software design. There’s a huge difference between using just the electricity you need to automate something versus stretching the entire electricity supply to cram everything into something that replaces human interaction and human capability.