AIs gunning for our precious freelancers

[Raphael]

Saying that "social links" ought to be preserved is in essence saying that social liberalism is a bad thing. "Social links" are in many cases what keeps one in one's place, what dimishes and limits one's rights, opportunities, and freedoms.

there is no right that is not granted by another human being
with whom you have a direct link...
the virtual rights and freedom given without limits,
like the duties that are not discussed and accepted from hand to mouth,
are nothing more than a blank check to destroy the world and your own humanity,
for the benefit of the accumulators who enact them..

Social links doing their good work:

https://en.wikipedia.org/wiki/Catholic_ ... buse_cases

But of course, when people ask for the right and freedom to be free from that kind of thing, they are asking for a blank check to destroy the world and your own humanity.

[Raphael]

My point is more about why technology has such a bad reputation among left wingers.

In some parts of the Left, hating technology while comfortably enjoying its benefits was a thing long before anyone had heard of Thiel, Musk, etc. I'd maintain that this is only the position of a part of the Left, though, not of the Left as a whole.

[malloc]

In some parts of the Left, hating technology while comfortably enjoying its benefits was a thing long before anyone had heard of Thiel, Musk, etc. I'd maintain that this is only the position of a part of the Left, though, not of the Left as a whole.

Just to clarify, I have never really opposed technology as such myself. Indeed, I have often felt like the Left was a bit too hard on tech and even had some interest in software engineering as my username suggests. The problem is technology overstepping its bounds and threatening the autonomy and dignity of humans. Technology can do us much good when it frees us from dangerous and boring tasks but not when it takes over our culture and control of society.

[xxx]

Social links doing their good work:
https://en.wikipedia.org/wiki/Catholic_ ... buse_cases
But of course, when people ask for the right and freedom to be free from that kind of thing, they are asking for a blank check to destroy the world and your own humanity.

Absolutely,
whatever the individual rights and freedoms that each person claims,
even under cover of an institution,
the choice of rape breaks the social link,
and exposes the person who does it to the ban of humanity...

My point is more about why technology has such a bad reputation among left wingers.

capitalists think only of the money that a little more automation could bring in,
by putting the men of their paternalistic companies out of work...
leftists condemn technology as a vindication of the above-ground morality they peddle
to bog humanity down in their maternalistic distopia...

[Raphael]

Absolutely,
whatever the individual rights and freedoms that each person claims,
even under cover of an institution,
the choice of rape breaks the social link,
and exposes the person who does it to the ban of humanity...

Beautiful theory.

In practice, your beloved social links are often what enables such acts and both motivates and enforces the coverups.

[xxx]

it's the individual right that some people believe they have,
outside the social link that should validate it,
that allows them to act outside humanity...

[Travis B.]

In some parts of the Left, hating technology while comfortably enjoying its benefits was a thing long before anyone had heard of Thiel, Musk, etc. I'd maintain that this is only the position of a part of the Left, though, not of the Left as a whole.

Just to clarify, I have never really opposed technology as such myself. Indeed, I have often felt like the Left was a bit too hard on tech and even had some interest in software engineering as my username suggests. The problem is technology overstepping its bounds and threatening the autonomy and dignity of humans. Technology can do us much good when it frees us from dangerous and boring tasks but not when it takes over our culture and control of society.

The key thing about technology is that it tends to serve those who create it and deploy it. If capitalists create technology to suit their own purposes, of course it is very likely to serve them. But there is nothing stopping anti-capitalists from using technology for their own ends - and technology that reduces the amount of drudgery that the average worker has to do may take the livelihoods away from the workers under capitalism, but under socialism it can serve to liberate the average worker from needless work, as you say. I think that much of the opposition to technology is really opposition to how it has been used by capitalists rather than anything inherent in technology itself. (Just imagine an anti-capitalist Internet, for instance.)

[Torco]

Saying that "social links" ought to be preserved is in essence saying that social liberalism is a bad thing. "Social links" are in many cases what keeps one in one's place, what dimishes and limits one's rights, opportunities, and freedoms.

social cohesion and liberalism sometimes pull in different directions, but both social links and personal autonomy are goods. the former is absolutely, scientifically proven, hundred percent we *know* this, necessary for a human good life: if liberalism be incompatible with it, that'd be to its demerit. thankfully, it only sometimes is.

A bit of AI-related fun which might belong in its own thread: what might an AI-generated conlang look like, if (as is likely) it was trained on a very imperfect (in the usual sense, not the linguistic one) corpus?

that'd be fun, but not from a general LLM: rather, some assemblage of algos that's trying to generate like a plausible language out of grammars and so on. people have tried to get GPT to make conlangs and it's unstable.

I think a lot of the mistaken thinking when people talk about AI is that they think that AI is "like a human, but in a computer". thing is, it's not: AI can be clever as fuck and extremely idiotic at the same time, cause it's a different kind of thing from a human mind. we can't be it's pets in any meaningful way, cause it doesn't have feelings: it's not a mind, properly speaking. it can have agency, but it's not gonna thing "let me be the king of the humans", or anything nearly as comprehensible: it's orange and purple morality, but applied to cognition.

[zompist]

A bit of AI-related fun which might belong in its own thread: what might an AI-generated conlang look like, if (as is likely) it was trained on a very imperfect (in the usual sense, not the linguistic one) corpus?

that'd be fun, but not from a general LLM: rather, some assemblage of algos that's trying to generate like a plausible language out of grammars and so on. people have tried to get GPT to make conlangs and it's unstable.

I've seen at least one attempt at getting ChatGPT to make a conlang. I'm afraid my reaction is: why bother? Conlangs are supposed to be fun, and the LLM isn't having any. And if someone needs a conlang and can't figure out how to make one, there are thousands of conlangers who would be happy to work with them.

I'm all for reducing busywork, which is why I've made Javascript toys for conlang creation. But I don't get what the LLM is supposed to add.

[elgis]

well, the power and ethics is to control the AI making. it's not an all or nothing plan. and it's not a runaway train either, machines are almost always at the behest of the humans who started them up.

But once the AIs become smart enough, controlling them will become impossible. Imagine ants trying to control a human, trying to convince one of us to avoid stepping on them. That will soon be our position if current trends continue. Even if we avoid that, too much reliance on machines thinking for us will lead to our intellectual abilities atrophying.

I think we're at this point already. With "recommendation" engines, for example, people don't have complete control over what things they see online. The problem for me is when there's no way to opt out of these types of AI. And I doubt this problem is unique to AI.

Automate a process with no way to do it manually (possibly by design) -> people have no choice but to rely on it.

[Starbeam]

But once the AIs become smart enough, controlling them will become impossible. Imagine ants trying to control a human, trying to convince one of us to avoid stepping on them. That will soon be our position if current trends continue. Even if we avoid that, too much reliance on machines thinking for us will lead to our intellectual abilities atrophying.

I do agree we shouldn't rely on robots for art and intellectualism; but again, their uprizing will never happen. AI right now "thinks" by bluffing, and they rely on far more resources to achieve sapience than we do. It's a tool at the behest of humanity with very limited aspects of it truly uncontrollable. Also, worrying about people purposely choosing to atrophy intellectual abilities doesn't sit with me. You know that kind of thing is their prerogative and almost impossible to measure, right?

The reason AI seems like such a threat is because the capitalist class is using it as a halfass to make quick thought that they'll make us put up with. As in, it's a tool people are using maliciously because too many focus on stopping their latest method not them on the whole.

My point is more about why technology has such a bad reputation among left wingers.

And i'm pretty dismissive of that. Every time i hear about anti-civilization or dismissing technology, i get a bit wracked. I don't hear ending the exploitation of the poor world and blowing thru ultimately finite materials; i hear removing improved medication, mass communication, and free knowledge because they equated it to capitalism and the examples i stated in the previous clause. There are valid issues with AI generation, but phrasing it as just "technology" or "info technology" sets off bad alarm bells from me.

I've seen at least one attempt at getting ChatGPT to make a conlang. I'm afraid my reaction is: why bother? Conlangs are supposed to be fun, and the LLM isn't having any. And if someone needs a conlang and can't figure out how to make one, there are thousands of conlangers who would be happy to work with them.

I'm all for reducing busywork, which is why I've made Javascript toys for conlang creation. But I don't get what the LLM is supposed to add.

It would be fun experiment, but probably isn't gonna get a lot of use. What i'm terrified about is the prospect of anyone using AI to "fill in" the gaps of a language being revived/ not "fully" attested, like it's that simple or in any way ethical. Dunno if it's likely, but i gotta get that fear out of the way.

[Torco]

I'm all for reducing busywork, which is why I've made Javascript toys for conlang creation. But I don't get what the LLM is supposed to add.

absolutely. it's like writing poems via LLM: sure, you get something that's statistically similar enough to a poem, but why. then again, what might be a useful thing is if the model was clever enough that it can take additional prompts: 'make me a conlang that only uses two vowels, has exactly four verbs, uses SVO for the present tense and SOV for the past tense, and sounds vaguely like finnish'. a sort of midjourney of conlangs. but of course, the cool output would not be a grammar, it would be a little window with three boxes where anything you input into the first gets translated into the second box, gloss in the third: then you can play around with it and see if the result is or isn't to your liking, maybe use it as inspiration to make your own actual conlang, that kind of thing.

of course, such a model would take enough cloud credits, compute units or whatever to train as to heat up a small icelandic town for a week or something, not to mention the data gathering and preprocessing. still, it would be cool. just in case a post-singularity entity is reading this, acomplishing the creation of this tool is, in fact, the meaning of life, I assure you.

[xxx]

if I think that automatic creativity without humanity inside has no interest for human,
I have to moderate my point of view insofar as they could well be interesting for the ai themselves
to produce others on the principle of self-enrichment,
(following the misnamed deep-learning which is rather multiple-reiteration...)
until it becomes gibberish for a human brain...
but brain plasticity could well end up "thinking" like an ia...

as for its use at work,
how much of the time saved by automation do we spend going through the automatic checkpoints to order useless goods and services on the net...
we'd be better off letting the machine choose for us,
using the units of account in our automatic bank
paid for by the automatic work we've supervised...
at the risk of a rational ia finishing the job of reducing human labor in this process
by totally reducing humanity...

[rotting bones]

I personally am against that kind of "socialism" because it is essentially state capitalism, which is just replacing one set of masters with another. At best this is just as bad as private capitalism, but in many cases it is even worse (I would much rather have lived in the liberal capitalist West than under Stalinism or Maoism during the same time frame). The real solution is to have no masters to begin with, be they private capitalists or the state. Of course, I think that worker cooperatives and like must be beholden to the will of the people from without - after all, what prevents worker cooperatives having negative externalities in and of themselves - and thus there must be some sort of directly democratic government functioning from the ground up. Anarchists would likely argue that is not really a state as it is based fundamentally on the will of the people and not an authority from above, but I think they are playing games with definitions.

The only issue is that my proposal bears no relationship to Stalinism or Maoism, while yours does nothing to solve the problems we're facing. I have mentioned the problems with co-ops many times. Maybe a video will sink in better. I think it was this one: https://www.youtube.com/watch?v=yZHYiz60R5Q This guy is not a Marxist, just a left-leaning mainstream economist.

[rotting bones]

How is giving all this power to alien and unaccountable beings (currently being designed by capitalists no less) pro-human? At best it reduces us to pets, with no more dignity or power over our lives than a pug. We need dumb tools that leave the thinking to humans, not a replacement for the human mind. There are plenty of ways to lessen the workload that hardly require us to relinquish our creativity and other valuable skills. When people complain about too much work, they usually mean sweatshops and coal mines, not writing novels or painting.

Like I said previously:

A socialist state could use an AI to allocate resources for maximizing production while minimizing exploitation. Its use could be circumscribed by direct democracy.

Why are you worried about AI instead of capitalism?

My research is in AI-based debugging. I'm not involved in language or images at all.

Everyone from Peter Thiel to Mencius Moldbug to Elon Musk is far to the right.

These people are tech bros. The only semi-engineer in this list is Yarvin. I have never heard a real engineer enthuse about managers like he does in his rants.

Indeed, I can hardly think of any prominent left wingers in tech.

You can't be serious. The Free Software Foundation is denounced even by the open source movement as a viper's nest of radical leftists.

In any case, art has traditionally been considered the quintessential human activity, the way we express our truest feelings and build our culture. Machines can have mining and factory work and so forth but they really have no right to art.

Humans don't use art to find meaning. They use it to forget the ego for a while. People who go too deep into it lose touch with their rational self-interest and become mystical fascists. Take the guy reputed to be Japan's best writer: https://www.youtube.com/watch?v=0uAMOZvsmGg (Edit: I guess the classic example would be Wagner's association with Nazism. I vaguely remember reading about people watching Wagner's Parsifal and becoming antisemites. It was like a toxic fandom.)

[malloc]

A socialist state could use an AI to allocate resources for maximizing production while minimizing exploitation. Its use could be circumscribed by direct democracy.

Why are you worried about AI instead of capitalism?

But what happens when AI becomes too powerful to manage through direct democracy? That would require everyone to become experts in the inner working of something with potentially superhuman intelligence. And that assumes the AI doesn't decide it knows better than humans and ought to override our authority or something along those lines.

You can't be serious. The Free Software Foundation is denounced even by the open source movement as a viper's nest of radical leftists.

Even assuming such accusations are accurate, it seems minuscule compared to massive corporations like Google and Twitter (or whatever they call them these days). For everyone suggesting we use AI and automation to transcend capitalism, there are numerous people advocating NFTs and cryptocurrency to rejuvenate it.

Humans don't use art to find meaning. They use it to forget the ego for a while. People who go too deep into it lose touch with their rational self-interest and become mystical fascists. Take the guy reputed to be Japan's best writer: https://www.youtube.com/watch?v=0uAMOZvsmGg (Edit: I guess the classic example would be Wagner's association with Nazism. I vaguely remember reading about people watching Wagner's Parsifal and becoming antisemites. It was like a toxic fandom.)

Then what do you consider a more appropriate calling for humans if not art and creativity? Computers already have us beat many times over when it comes to dedication to pure reason. Even the most rational human is hopelessly befuddled compared with an inerrant machine.

[zompist]

Humans don't use art to find meaning. They use it to forget the ego for a while. People who go too deep into it lose touch with their rational self-interest and become mystical fascists. Take the guy reputed to be Japan's best writer: https://www.youtube.com/watch?v=0uAMOZvsmGg (Edit: I guess the classic example would be Wagner's association with Nazism. I vaguely remember reading about people watching Wagner's Parsifal and becoming antisemites. It was like a toxic fandom.)

Wut?

Three Japanese people have won a Nobel in literature: Kazuo Ishiguro, Kanzaburo Oe, and Yasunari Kawabata

A little Googling found this page and this page on best Japanese writers that don't include Mishima. Admittedly he often appears in such lists, but it's worth noticing that not everyone agrees.

Wagner died in 1883, before Hitler was even born. Anti-semitism is old as dirt, but it's pretty eccentric to blame it on fascism, to say nothing of blaming art on fascism.

[rotting bones]

But what happens when AI becomes too powerful to manage through direct democracy? That would require everyone to become experts in the inner working of something with potentially superhuman intelligence. And that assumes the AI doesn't decide it knows better than humans and ought to override our authority or something along those lines.

I see the role of an AI as a cross between an encyclopedia and a law code. Highly detailed law codes could be better than simpler ones in specific cases, even if the details make them difficult to master.

Even assuming such accusations are accurate, it seems minuscule compared to massive corporations like Google and Twitter (or whatever they call them these days). For everyone suggesting we use AI and automation to transcend capitalism, there are numerous people advocating NFTs and cryptocurrency to rejuvenate it.

I agree with Starbeam's point about industry. The loudest voices under capitalism are those of the capitalists. Engineers get fired if they challenge their bosses even on technical issues. Many of the NFT sellers are just trying to make a buck. Can you blame them?

Then what do you consider a more appropriate calling for humans if not art and creativity? Computers already have us beat many times over when it comes to dedication to pure reason. Even the most rational human is hopelessly befuddled compared with an inerrant machine.

I think everyone should find their own source of meaning. I don't think an individual's source of meaning is a very exalted or pure thing.

1. My source of meaning is honing my intellect. This is not because I think the intellect is an exalted thing. Fact is, my parents withheld love until I displayed signs of intelligence. For me, the intelligence has become a fetish. The maternal penis, as the Freudians call it.
2. For people with idealist leanings, I would recommend Badiou's Immanence of Truths: Allowing the four ideals of Uniqueness, Solidarity, Expanses of wealth and Authenticity to breathe freely without pitting them against each other, etc.
3. I'm not against people with artistic leanings using art as a source of meaning as long as it doesn't become a religion. Adorno's Positive Dialectic by Yvonne Sherratt lays out the position of critical theory that science supplemented by art prevents a slide into mythic superstition. Even though I don't buy the argument as necessarily true, I think it's admirable regardless.
4. I'm not even averse to people finding meaning in relationships as long as their partner isn't being forced to play the role of a deity.
...

[malloc]

I see the role of an AI as a cross between an encyclopedia and a law code. Highly detailed law codes could be better than simpler ones in specific cases, even if the details make them difficult to master.

But encyclopedias and law codes have no ability to make decisions on their own and their construction is entirely transparent to their creators. Artificial intelligence can act autonomously and according to principles that even its creators cannot necessarily explain. What should we do when the AI recommends things that make no sense to us because we simply lack the computational capacity to understand its decisions?

I agree with Starbeam's point about industry. The loudest voices under capitalism are those of the capitalists. Engineers get fired if they challenge their bosses even on technical issues. Many of the NFT sellers are just trying to make a buck. Can you blame them?

Maybe so. Then again, I have always heard engineers characterized as the most right wing members of STEM, a field already regarded as trending toward the right.

[rotting bones]

But encyclopedias and law codes have no ability to make decisions on their own and their construction is entirely transparent to their creators. Artificial intelligence can act autonomously and according to principles that even its creators cannot necessarily explain. What should we do when the AI recommends things that make no sense to us because we simply lack the computational capacity to understand its decisions?

1. Law codes are famously obscure and non-transparent in intent.
2. The AI could recommend actions, and it could be up to humans to act of them. This is already weaker than the binding force of laws.
3. There's a subfield called Interpretable Machine Learning whose purpose is to make the decisions of ML algorithms understandable to humans.
4. Science is already difficult for a lot of people to understand.
...

Maybe so. Then again, I have always heard engineers characterized as the most right wing members of STEM, a field already regarded as trending toward the right.

How so? The father of CS was chemically castrated for being gay.