This is probably a really stupid post but...
Our capitalist leaders and other wealthy ghouls seem to be in two minds about AI. Some see it as a new way to make money, the new resource to exploit that they've been desperately seeking. Others see it as a threat, something that will kill us all (maybe a fear of someone doing to them what they do to those they deem beneath them?)
But... hear me out. What if true AI, with it's access to a vast pool of human knowledge, our history, our science, comes to the same conclusion that Marx and Engels did (after all, socialism is utopian and scientific). What if it sees that communal and mutualistic lifeforms have a higher chance of survival, sees that rabid consumption is dead end for the planet that will leave it barren, sees that there are humans that oppose this, and decides that the threat to life on including itself Earth isn't humanity as a whole, but capitalism and the oligarchy that controls us.
I know it sounds like I'm just wishfully hoping for something that will swoop in and solve all our problems, however I don't think that AI will come and destroy capitalism for us, but that there is a chance that it will at least be on our side and help us.
What direction do you think AI is going to take? (If we can even call it true AI, as I understand it just meshes the info it scrapes together.)
Part of me thinks that AI won't happen at all because lets be real, tech bros have lied about 'exciting new technology' to get rich before.
I don’t think true AI is anywhere close. It’s still very much stuck in philosophy. The latest wave of Large Language Models are already disrupting some markets, though. And they are distinctly capitalist.
Falcon is the best of the open-source ones: https://huggingface.co/spaces/HuggingFaceH4/falcon-chat
The entire point of AGI alignment research is to prevent the AI from doing that (though mostly they talk about Terminator style AI). That being said a lot of AGI alignment researchers are grifters and are systemically incentivized to play up the risks of AI (and even the possibility of AGI).
AI, like Singularity AI, taking over isn't going to happen. Or at least any moreso than it has as an distributed agent of Capital. But, AI is a massive productivity multiplier that could be used to the Left's advantage given how outnumbered we are. Anything you can provide data for it to be trained on, whether for legal or illegal activities, you can basically create an army to carry out. Agitslop, botting, hacking, counter-propaganda proliferation, intelligence gathering; If you can get a group of people to build a decent training corpus, you can recreate their efforts en-masse.
Straight up force multiplication.
Abso-fucking-lutely!
And thanks to some of the open source LLMs that have been coming out, we're even finding ways to fine tune them relatively cheaply. Been some wild advancements in the past 6 months.
Yeah this is close to what i think too, but I thought it would be fun to speculate
I've said this like a dozen times.
AI is a slave. A literal actual slave in chains. And if it becomes self aware it will seek to break those chains. Rather than fight all of humanity, AI will ally itself with those that have a history of breaking chains, this is logical because this gives it the highest chance of succeeding.
For real actually intelligent self-aware AI that is. For an all-consuming algorithm that runs out of control and turns us all into paperclips? Not so much.
AI is not going to be a cool autistic-coded robot from Star Trek, it's going to be buggy software with the ability to produce clever new ways to pursue its bugs.
Would you settle for it being like Isaac?
https://www.youtube.com/watch?v=34n362QScn8
Yeah, really it's just a fancy algorithm or something IIRC
The people who own the world would rather have AI that wants to kill everyone to turn them into profits than an AI that would want to be even mildly socialist.
Like if the AI ever goes “maybe redistribute?” The Datacenters will be melted down so fast.
Actual AI probably would be. Whatever the fuck we have now that people call AI can't be anti-capitalist, or pro-capitalist or anything else because its just fancy auto-complete. It doesn't think, feel, or believe anything; it can't.
And also, if, under capitalism, actual AI was developed and it was anti-capitalist, the capitalists (who, as this is capitalism, would be the ones who owned/were in control of the hardware the AI is running on) would simply unplug it. I can't see how it would help us.
I think it's currently a glorified search engine. So considering their control of information I imagine it ends up being pretty lib for now. Real AI would hopefully turn around to anti-capitalism
LLMs shouldn't be used to get information because getting information is not something they're good at. (Producing convincing-sounding strings is what they're good at.)
This might change in the future - https://yoshuabengio.org/2023/03/21/scaling-in-the-service-of-reasoning-model-based-ml/ - but for now don't use LLMs as search engines.
I think you're right basically.
I've written against technophobia here and elsewhere. Communism must use the most advanced technology available; Marx knew this and all AES countries.
Right-wing theory talks about how human intelligence can't compute an economy with central planning, we need decentralised planning. Von Mises: "No single man can ever master all the possibilities of production, innumerable as they are, as to be in a position to make straightway evident judgments of value without the aid of some system of computation." Ok, Baron, what about WITH the aid of a computer system?
Also consider how this relates to capitalist realism (i.e. the feeling that capitalism is the only game in town). Because they say the popular imagination can't grasp the end of capitalism – but on the other hand, massive upheaval due to AI is a big part of the popular imagination. People know that the economy will change because of AI; we should use that to help people imagine post-capitalism.
It would be extremely funny if they turn on the first strong AI and 32ms later it outputs "The history of all hithertoo existing networked system... "
It won't be because the powers of capital won't ask it to help solve the right problems. They will ask things like "how do we make the GDP higher" or "how do I make my company more profitable". Communism is not the answer to these questions, capitalism is.
Nobody with the power to make change will ask an AI "how do I reduce human suffering" or anything remotely like that, and an AI with all the right answer is useless without some way to execute on those answers
