Farming took half the population, so we mechanized that, and told ex-farmhands they still needed to work.
Laundry was arduous and dangerous, so we mechanized that, and told ex-washerwomen they still needed to work.
Dishwashing constantly broke things, so we mechanized that, and told ex-scullery maids they still needed to work.
Math is eminently fuck-up-able, so we mechanized that, and… actually accountants with spreadsheets just took on a bunch more work.
Point is: art is damn near the last thing we’ve mechanized. And it happened almost by accident. If we knew how to spend a billion dollars and get Asimov robots instead, we’d do that, no question. Even the ultracapitalist douchebags would pick that. Christ, it’s not like artists have much money to take.
Labor-saving technology is all around you. It just hasn’t been allowed to save labor. Some new tech absolves humanity of such-and-such drudgework, and the rich pricks in charge tell the ex-drudgeworkers, better hustle harder. Maybe the tech isn’t the problem.
Automation is a great tool that isn’t based on stolen work, so could be LLMs. But the issue here is that we’ve had corporations going to extreme lengths to enforce copyright but now they want to scrape anything and everything, regardless of copyrights? And it’s to replace artists? Fuck that noise.
LLMs/generative AI could create clean datasets by paying people to generate curated data and remunerate the artists that the AI is based on. That would be a start and miles more ethical than it is now.
That’s disingenuous calling AI slop mechanized art. It isn’t. Capitalists have proven time and time again that they would kill people if that means more profit. So your point about Asimov robots is moot.
That’s kettle logic. Individually defensible points only line up at the conclusion.
It should be done ethically! Even though it’s fundamentally evil for replacing artists.
It’s never art! As if anything but art could replace the creative arts.
Murderous capitalists should target monotony first! My guy: they did.
The point about Asimov robots is a direct response to your initial demands. You want them to do that. Guess what: so do they. They’ve been trying, for a century. We don’t know how. But we stumbled into doing this, and now we know how to do this, and it’s not gonna go away.
Nor is it going to replace artists. You still need people, to make things, with this tool… if you choose to use this tool, to make things. CGI’s not going anywhere. Pointing a camera at actors certainly isn’t going anywhere. Traditional animation’s fucked, but blame CGI for that. Zero people will be prevented from expressing themselves through the medium of drawing, thanks to the hot anime girl machine. But it lets people fancy-up whatever they’ve done.
This is the tech that will let 2D animators draw one complete frame per second and get smooth interpolation. If the machine guesses wrong, oh well, gotta draw some more frames the hard way. This is the tech that will let any writer film their trunk script, exactly how they see it in their head, provided they have a camera and some action figures. They’ll have to do all the voices themselves, but it’ll sound like multiple people, perfectly mimicking his mediocre performance. This is the tech that will let high-effort “photorealistic” CGI bridge the gap to alarmingly convincing verisimilitude, using details we instantly notice, but cannot put into words. If we knew how to do it the hard way, we already would.
On a more constructive bent, because apparently I have a compulsion:
Nothing stops companies from paying people to build their dataset, but they don’t because they are greedy as fuck.
Blaming how the model was built doesn’t make sense if you think having a model is evil. If you expect this “nifty” technology will destroy the creative industry - what difference does the dataset make?
Open-source nerds would just make their own for free. There’s a ton of CC-licensed content, even if you exclude the people going 'well I didn’t know it would be used for that!’ There will be a model for which nobody is paid, because some dorks will stone-soup their own libre / gratis version. When that happens - will it change your opinion of this basically identical situation?
Except that the technology at the current level is used to not pay artists, actors and many creatives their fair share.
What would it look like, if it looked like what you wanted? What possible future are you alluding to? Describe what you’d consider an ethical use of an ethically-derived model. Maybe keep it in mind, when you see other commenters here railing against the technology, as a concept.
The objections to AI image gens, training sets containing stolen data, etc. all apply to LLMs that provide coding help. AI web crawlers search through git repositories compiling massive training sets of code, to train LLMs.
Art is meant to be an expression of the self and a form of communication. It’s therapeutic, it’s liberating, it’s healthy and good. We make art to make and keep us human. Chatbot art doesn’t help us, and in fact it makes us worse - less human. You’re depriving yourself of enrichment when you use a chatbot for art.
Code is mechanical and functional, not really artistic. I suppose you can make artistic code, but coders aren’t doing that (maybe they should, maybe code should become art, but for now it isn’t and I think that’s a different conversation). They’re just using tools to perform a task. It was always soulless, so nothing is lost.
Art is also functional. Specifically, paid opportunities for art perform some type of function. Not all art is museum type contemplative work or the highest level of self expression. Some art, its purpose is to serve as a banner on the side of a web page notifying people of a new feature. That isn’t really enriching to create. It’s a rather low form of self expression, similar to code created to be functional.
I think you’re also underestimating AI image gens as a form of self expression. Obviously it’s more freeing to be able to draw or paint or create a thing yourself. But people often lack the prerequisite skills to create the thing they have in their mind. I often see career artists compare their work and style from years ago to their works today, showing off extreme improvement - meaning that even talented artists sometimes lack the skills necessary to create the “perfect” version of what they had in their mind.
With LLMs, you can get quite specific - not just “draw me in a Studio Ghibli style,” but meticulously describing a scene and visual style - and it will render it. There is still creative satisfaction in that process, like how a movie director tells the actors how to accomplish a scene but doesn’t actually play a role in the film themselves.
But people often lack the prerequisite skills to create the thing they have in their mind.
And they will always lack those skills if they never practice!
Furthermore, art isn’t just functionally putting creations into the world, it’s also the act of creation. There’s a feeling of creation that comes from creating art, it’s about the journey and not just the destination.
Having a chatbot do it for you isn’t the same.
There is still creative satisfaction in that process, like how a movie director tells the actors how to accomplish a scene but doesn’t actually play a role in the film themselves.
Many actors do not want to be directors, many directors do not want to be actors. Those are just different things.
Even if you want to compare prompting LLMs with directing, that still means that people are deprived of acting. They’re missing out on feeling and experiencing the act of artistic expression by outsourcing it to a chatbot.
“i am fine with stolen labor because it wasn’t mine. My coworkers are falling behind because they have ethics and don’t suck corporate cock but instead understand the value in humanity and life itself.”
Then most likely you will start falling behind… perhaps in two years, as it won’t be as noticable quickly, but there will be an effect in the long term.
I use gpt to prototype out some Ansible code. I feel AI slop is just fine for that; and I can keep my brain freer of YAML and Ansible, which saves me from alcoholism and therapy later.
You probably create AI slop and present it proudly to people.
AI should replace dumb monotonous shit, not creative arts.
We did.
Farming took half the population, so we mechanized that, and told ex-farmhands they still needed to work.
Laundry was arduous and dangerous, so we mechanized that, and told ex-washerwomen they still needed to work.
Dishwashing constantly broke things, so we mechanized that, and told ex-scullery maids they still needed to work.
Math is eminently fuck-up-able, so we mechanized that, and… actually accountants with spreadsheets just took on a bunch more work.
Point is: art is damn near the last thing we’ve mechanized. And it happened almost by accident. If we knew how to spend a billion dollars and get Asimov robots instead, we’d do that, no question. Even the ultracapitalist douchebags would pick that. Christ, it’s not like artists have much money to take.
Labor-saving technology is all around you. It just hasn’t been allowed to save labor. Some new tech absolves humanity of such-and-such drudgework, and the rich pricks in charge tell the ex-drudgeworkers, better hustle harder. Maybe the tech isn’t the problem.
Automation is a great tool that isn’t based on stolen work, so could be LLMs. But the issue here is that we’ve had corporations going to extreme lengths to enforce copyright but now they want to scrape anything and everything, regardless of copyrights? And it’s to replace artists? Fuck that noise.
LLMs/generative AI could create clean datasets by paying people to generate curated data and remunerate the artists that the AI is based on. That would be a start and miles more ethical than it is now.
That’s disingenuous calling AI slop mechanized art. It isn’t. Capitalists have proven time and time again that they would kill people if that means more profit. So your point about Asimov robots is moot.
That’s kettle logic. Individually defensible points only line up at the conclusion.
It should be done ethically! Even though it’s fundamentally evil for replacing artists.
It’s never art! As if anything but art could replace the creative arts.
Murderous capitalists should target monotony first! My guy: they did.
The point about Asimov robots is a direct response to your initial demands. You want them to do that. Guess what: so do they. They’ve been trying, for a century. We don’t know how. But we stumbled into doing this, and now we know how to do this, and it’s not gonna go away.
Nor is it going to replace artists. You still need people, to make things, with this tool… if you choose to use this tool, to make things. CGI’s not going anywhere. Pointing a camera at actors certainly isn’t going anywhere. Traditional animation’s fucked, but blame CGI for that. Zero people will be prevented from expressing themselves through the medium of drawing, thanks to the hot anime girl machine. But it lets people fancy-up whatever they’ve done.
This is the tech that will let 2D animators draw one complete frame per second and get smooth interpolation. If the machine guesses wrong, oh well, gotta draw some more frames the hard way. This is the tech that will let any writer film their trunk script, exactly how they see it in their head, provided they have a camera and some action figures. They’ll have to do all the voices themselves, but it’ll sound like multiple people, perfectly mimicking his mediocre performance. This is the tech that will let high-effort “photorealistic” CGI bridge the gap to alarmingly convincing verisimilitude, using details we instantly notice, but cannot put into words. If we knew how to do it the hard way, we already would.
Except that the technology at the current level is used to not pay artists, actors and many creatives their fair share. So your point is moot.
And even worst, the generative models use the work of these same artists to build their data sets to replace them, without even acknowledging them.
The tech is pretty nifty and useful, but is used to avoid paying the fair price of the creatives’ work.
Nothing stops companies from paying people to build their dataset, but they don’t because they are greedy as fuck.
On a more constructive bent, because apparently I have a compulsion:
Blaming how the model was built doesn’t make sense if you think having a model is evil. If you expect this “nifty” technology will destroy the creative industry - what difference does the dataset make?
Open-source nerds would just make their own for free. There’s a ton of CC-licensed content, even if you exclude the people going 'well I didn’t know it would be used for that!’ There will be a model for which nobody is paid, because some dorks will stone-soup their own libre / gratis version. When that happens - will it change your opinion of this basically identical situation?
What would it look like, if it looked like what you wanted? What possible future are you alluding to? Describe what you’d consider an ethical use of an ethically-derived model. Maybe keep it in mind, when you see other commenters here railing against the technology, as a concept.
You can’t just say perchance.
The model itself isn’t evil, how it is used is.
Ethical use : pay artists to build your dataset and a royalty on generated art.
The moment this is the case, I won’t have any qualm about using it. Artists still get paid for their creative work throughout the process.
It’d be great if your responses had anything to do with the words that I wrote.
deleted by creator
That’s what the OP is about, so…
Has AI made you unable to read?
The objections to AI image gens, training sets containing stolen data, etc. all apply to LLMs that provide coding help. AI web crawlers search through git repositories compiling massive training sets of code, to train LLMs.
deleted by creator
But your opinion is off topic.
deleted by creator
Code and art are just different things.
Art is meant to be an expression of the self and a form of communication. It’s therapeutic, it’s liberating, it’s healthy and good. We make art to make and keep us human. Chatbot art doesn’t help us, and in fact it makes us worse - less human. You’re depriving yourself of enrichment when you use a chatbot for art.
Code is mechanical and functional, not really artistic. I suppose you can make artistic code, but coders aren’t doing that (maybe they should, maybe code should become art, but for now it isn’t and I think that’s a different conversation). They’re just using tools to perform a task. It was always soulless, so nothing is lost.
Art is also functional. Specifically, paid opportunities for art perform some type of function. Not all art is museum type contemplative work or the highest level of self expression. Some art, its purpose is to serve as a banner on the side of a web page notifying people of a new feature. That isn’t really enriching to create. It’s a rather low form of self expression, similar to code created to be functional.
I think you’re also underestimating AI image gens as a form of self expression. Obviously it’s more freeing to be able to draw or paint or create a thing yourself. But people often lack the prerequisite skills to create the thing they have in their mind. I often see career artists compare their work and style from years ago to their works today, showing off extreme improvement - meaning that even talented artists sometimes lack the skills necessary to create the “perfect” version of what they had in their mind.
With LLMs, you can get quite specific - not just “draw me in a Studio Ghibli style,” but meticulously describing a scene and visual style - and it will render it. There is still creative satisfaction in that process, like how a movie director tells the actors how to accomplish a scene but doesn’t actually play a role in the film themselves.
And they will always lack those skills if they never practice!
Furthermore, art isn’t just functionally putting creations into the world, it’s also the act of creation. There’s a feeling of creation that comes from creating art, it’s about the journey and not just the destination.
Having a chatbot do it for you isn’t the same.
Many actors do not want to be directors, many directors do not want to be actors. Those are just different things.
Even if you want to compare prompting LLMs with directing, that still means that people are deprived of acting. They’re missing out on feeling and experiencing the act of artistic expression by outsourcing it to a chatbot.
“i am fine with stolen labor because it wasn’t mine. My coworkers are falling behind because they have ethics and don’t suck corporate cock but instead understand the value in humanity and life itself.”
deleted by creator
Then most likely you will start falling behind… perhaps in two years, as it won’t be as noticable quickly, but there will be an effect in the long term.
deleted by creator
I know senior devs who fell behind just because they use too much google.
This is demonstrably much worse.
deleted by creator
I use gpt to prototype out some Ansible code. I feel AI slop is just fine for that; and I can keep my brain freer of YAML and Ansible, which saves me from alcoholism and therapy later.