AIs gunning for our precious freelancers

[Raphael]

I don't think that abolishing capitalism would necessarily mean the end of AI. It might as well mean the switch from a state of affairs where new labor-saving devices mean "Oh no, a lot of us will lose our jobs!" to a state of affairs where new labor-saving devices mean "Yay, we all have to work less!"

[WeepingElf]

I don't think that abolishing capitalism would necessarily mean the end of AI. It might as well mean the switch from a state of affairs where new labor-saving devices mean "Oh no, a lot of us will lose our jobs!" to a state of affairs where new labor-saving devices mean "Yay, we all have to work less!"

Yes. While I was a democratic socialist in my youth, I now feel that the socialists are barking up the wrong tree. It is not a question of who owns the businesses, it is a question of what the businesses do. That's a question of regulation and legislation, and hence, democracy is the thing, not socialism. Indeed, I consider starting and running a business a legitimate way of self-realization, as long as you run the business in a way that doesn't harm other people or the environment.

[Raphael]

I don't think that abolishing capitalism would necessarily mean the end of AI. It might as well mean the switch from a state of affairs where new labor-saving devices mean "Oh no, a lot of us will lose our jobs!" to a state of affairs where new labor-saving devices mean "Yay, we all have to work less!"

Yes. While I was a democratic socialist in my youth, I now feel that the socialists are barking up the wrong tree. It is not a question of who owns the businesses, it is a question of what the businesses do.

Err, I think that's pretty much the opposite of my point.

[WeepingElf]

I don't think that abolishing capitalism would necessarily mean the end of AI. It might as well mean the switch from a state of affairs where new labor-saving devices mean "Oh no, a lot of us will lose our jobs!" to a state of affairs where new labor-saving devices mean "Yay, we all have to work less!"

Yes. While I was a democratic socialist in my youth, I now feel that the socialists are barking up the wrong tree. It is not a question of who owns the businesses, it is a question of what the businesses do.

Err, I think that's pretty much the opposite of my point.

Oh, so I utterly misunderstood you - but that doesn't diminish my point about socialism.

[Travis B.]

I don't think that abolishing capitalism would necessarily mean the end of AI. It might as well mean the switch from a state of affairs where new labor-saving devices mean "Oh no, a lot of us will lose our jobs!" to a state of affairs where new labor-saving devices mean "Yay, we all have to work less!"

Yes. While I was a democratic socialist in my youth, I now feel that the socialists are barking up the wrong tree. It is not a question of who owns the businesses, it is a question of what the businesses do. That's a question of regulation and legislation, and hence, democracy is the thing, not socialism. Indeed, I consider starting and running a business a legitimate way of self-realization, as long as you run the business in a way that doesn't harm other people or the environment.

To me actual socialism, as opposed to the state capitalism that is commonly confused with socialism, is all about making the economy and the world more democratic - and not in the Communists' fake "democratic" way either. Workers will not willingly vote themselves out of a job and subject themselves to poor working conditions or poor compensation. Of course, the interests of society as a whole also have to be taken into account, because there is nothing which, by itself, prevents a worker cooperative from polluting or overexploiting the environment any more than is the case with the current private capitalism. This is why I now support a directly democratic government which strongly regulates such worker cooperatives. It must be be directly democratic, as the current liberal democracy has failed to do so in very many cases and has been coopted by the capitalists, and only direct democracy will truly act in the interests of the population. Note that electoralism and parliamentarianism have shown themselves to have failed in this regard, and something like workers' councils (as they are called, even though I personally think it is a poor name) are needed for such direct democracy.

[Raphael]

and only direct democracy will truly act in the interests of the population.

Referendums seem to have had a very mixed track record recently. The country with the most direct democracy is Switzerland, which is not known for its progressivism.

[Travis B.]

and only direct democracy will truly act in the interests of the population.

Referendums seem to have had a very mixed track record recently. The country with the most direct democracy is Switzerland, which is not known for its progressivism.

I am personally against referenda for this reason - remember that dictators have often used plebiscites to legitimize themselves. But referenda are not the only sort of direct democracy by any means. I am much more for a workers' council model than a referenda model for this very reason.

[Raphael]

I am personally against referenda for this reason - remember that dictators have often used plebiscites to legitimize themselves. But referenda are not the only sort of direct democracy by any means. I am much more for a workers' council model than a referenda model for this very reason.

If it's done through some kind of council, how is that different from representative democracy?

[Travis B.]

I am personally against referenda for this reason - remember that dictators have often used plebiscites to legitimize themselves. But referenda are not the only sort of direct democracy by any means. I am much more for a workers' council model than a referenda model for this very reason.

If it's done through some kind of council, how is that different from representative democracy?

What a workers' council is is where you have groups of individuals, historically typically workplaces but not necessarily workplaces, who elect one (or more) of themselves to serve as delegates for themselves in a higher workers' council above them to communicate their wishes. Unlike in representative democracy, such delegates are only supposed to communicate decisions made by those who selected them, and are arbitrarily and immediately recallable by them (unlike in representative democracy, where recalls are generally a long drawn-out, arduous affair).

[malloc]

I don't think that abolishing capitalism would necessarily mean the end of AI. It might as well mean the switch from a state of affairs where new labor-saving devices mean "Oh no, a lot of us will lose our jobs!" to a state of affairs where new labor-saving devices mean "Yay, we all have to work less!"

Aside from the loss of jobs, one must also consider the risks of giving too much control of the economy to what are effectively black boxes. All existing tool operate according to predictable rules chosen by their creators. The hammer a carpenter can only do what the carpenter wants and makes no decisions of its own. But what happens when AI becomes powerful enough that we can no longer predict its behavior or understand the logic behind its decisions? It ceases to be a tool and becomes another form of unaccountable authority.

[Torco]

I don't think that abolishing capitalism would necessarily mean the end of AI. It might as well mean the switch from a state of affairs where new labor-saving devices mean "Oh no, a lot of us will lose our jobs!" to a state of affairs where new labor-saving devices mean "Yay, we all have to work less!"

granted, but AI is not the problem: AI is just a bunch of linear algebra you can use to cassify things into buckets, calculate how likely it is someone will like a certain book given that they like this other book, or turn prompts into images and videos by generalizing what videos are like already. eventually it will be able to do more things, such as impersonating a person, and automating more and more tasks: that's not bad, that's pretty fucking awesome! the problem is that *what is done with the AI*, just like *what is done with all other tools*, is decided by a special class of people and, because of the arbitrary rules of the economy, they can only do what will make themselves or other rich people even richer. that's the only task, ultimately, AI is ever applied to: and, yeah, *that* is the problem.

but if the economy was managed democratically... do you think people would vote themselves into starvation? do you think like, I don't know, your local bureau of economic affairs for your city or whatever, if that bureau was like a town hall where the guys and gals and enbies who live there vote and discuss proposals... do you think they'd vote for the "we all starve but Bob gets to buy whatever he wants even more than he already does" motion ? or would they vote for the "look, the robots really could grow all of the food, so how about we automate however many farms and we each get like 1/whateverth of the food."

like, fuck lenin and stalin and blabla, that's the *idea* of socialism: economic democracy. under it, AI would be pretty cool. the problem with AI is the problem with all technology: it's not that technology is bad, is that owners just get to do whatever they want with it. pollute, buy politicians, make life more difficult for others, all just so owner man gets more dosh.

Yes. While I was a democratic socialist in my youth, I now feel that the socialists are barking up the wrong tree. It is not a question of who owns the businesses, it is a question of what the businesses do.

Exactly, man! ownership is a social construct, no more. what matter is the rules! our current rule is "each business is its own absolute monarchy, subject to some laws and to not running out of money but ultimately it does what the owner wants: owner wants to do it? is it legal? we do it. perod". that's a shitty rule. other rules would be better.

Referendums seem to have had a very mixed track record recently. The country with the most direct democracy is Switzerland, which is not known for its progressivism.

but it has a stellar track record of acting in its own geopolitican interests, though.

the risks of giving too much control of the economy to what are effectively black boxes.

nah, that's not the danger of AI. if a model outputs something that's against the ownerman's interests, it will be shut down. the models are trained to maximize profit, ultimately, because the business itself is: AI requires training and utility functions you're trying to maximize, and the main source of loss function divergence between models and people is wherever those loss functions, or utility functions, are coming from.

[malloc]

nah, that's not the danger of AI. if a model outputs something that's against the ownerman's interests, it will be shut down. the models are trained to maximize profit, ultimately, because the business itself is: AI requires training and utility functions you're trying to maximize, and the main source of loss function divergence between models and people is wherever those loss functions, or utility functions, are coming from.

But plenty of AI enthusiasts are trying to create sentient AGI with superhuman intelligence. They have this goal they call the Singularity where AI advances to the point where it achieves godlike intellect and power. Apparently they are common enough in the tech industry that the acronym TESCREAL has been coined to describe them.

[Travis B.]

[snip]

I must say that I agree completely there ─ in a truly socialist society, AI would serve the workers rather than the capitalists or the apparatchiks because it would be the workers who would provide the training and the utility functions for the AI so the AI would serve them. AI would serve to reduce the amount of work the workers have to do while increasing what they reap from it, rather than reducing their pay and potentially putting them out of jobs.

[Torco]

But plenty of AI enthusiasts are trying to create sentient AGI with superhuman intelligence.

yee but that's such an ill-defined problem. ai agents are a thing, though. I'm pretty sure you can wreack a lot of havoc with quite ungeneral intelligence

AI would serve to reduce the amount of work the workers have to do while increasing what they reap from it, rather than reducing their pay and potentially putting them out of jobs.

that's the idea, anyway.

also... going back to our precious freelancers: might not the opposite be true? AI could automate the breaking down of tasks to smaller tasks, thus effectively doing away with managers: managers don't really *do* that much to begin with, in general the details and little challenges of a project get worked out by the five or six guys actually doing the thing. AI managers, everyone freelancers. *then* an AI could go haywire and cause big damage.

[zompist]

AI could automate the breaking down of tasks to smaller tasks, thus effectively doing away with managers: managers don't really *do* that much to begin with, in general the details and little challenges of a project get worked out by the five or six guys actually doing the thing.

That's the big irony, I think: executives think LLMs can replace creative work because their work is easily automated with LLMs. Most of their job is producing plausible-sounding gibberish.

That's probably not entirely fair, because the old-style capitalists actually had to know their business. But the whole idea of MBA-land is that a few years learning gibberish-production qualifies you to "manage" absolutely anything: if you actually know things you won't get promoted. They dumbed down the job so much that it's barely a job at all.

I hate to say it, but a Soviet-style commissar can also be easily automated. The avant-garde party is also mostly a matter of producing gibberish while being profoundly out of touch.

[elgis]

On the other hand, I would prefer to manage AI than for AI to manage me.

[Torco]

you, too can become a manager! just scan your mind, get a giftcard!

[xxx]

whatever the political regime, the individual is subject to a choice of society over which he has no control...
at most, they can keep their physical or mental little paradise, depending on the degree of control of the society in which they live...

no matter how intelligent we are, the global choice is always the law of the jungle...
the problem is that as technology takes us further away from the jungle, it borders on the complete destruction of the jungle...
which puts an end to all individual physical paradise in the medium term...

the new problem with a technology like AI is that it's the possibility of choice that eludes us...
and by taking away the use of our reasoning to make life choices, it's the very possibility of reasoning that's slipping away...
from there to think that it would close off the possibility of even preserving an autonomous mental paradise...



some with a sheepish or routine mind would perhaps be satisfied with such a world
where everything is decided by others and where you only have to follow an established order, it is perhaps the majority...

Brave New Word is a dystopia seen from here
but a happy utopia for the majority of its characters...
and so close to today...

[Torco]

some with a sheepish or routine mind would perhaps be satisfied with such a world
where everything is decided by others and where you only have to follow an established order, it is perhaps the majority...

subjugation isn't a new thing, though. graeber and wrenwrow have good references to accounts of what, for example, native americans thought of western society: they said much the same thing about people from the 1600s. already most people live paycheck to paycheck, one financial misstep away from entering the dalit class that is homeless people, spending the majority of their waking lives in the abject service of others, having their opinions and values shaped by enormously more skilled manipulators, liars and showmen. this is true now, this was true in tenth century japan, and it was true in the manors of the romans. sure, AI will be yet another tool of the ruling class, but so is steel or multimodal logistics.

[Raphael]

It might be interesting to compare and contrast today's real life debates about AI with the ones predicted by zompist's old story Rights for Robots:

http://zompist.com/robot.htm

In the real 2023, left-wingers are usually sceptical about or outright opposed to AI, while right-wingers who aren't rich capitalists seem to mostly ignore it - or at least I haven't heard much about their reactions to it - and right-wingers who are rich capitalists are simultaneously alarmist and enthusiastic about it.

In zompist's old story, stereotypical left-wing activist types are mainly concerned about AIs' rights being violated by human beings, and the people hostile to AIs seem to be populist right-wingers in the mold of classical "foreigners are taking our jobs"-groups.

Another thing to note is that in zompist's old story, all the serious research into AI seems to be done by major academic institutions, while in the real 2023, it seems to be done by major (and perhaps minor) tech corporations.

Finally, I'd note that none of the entities that get called "AIs" in the real 2023 seem to be nearly as smart as the AIs in zompist's old story.