I’ve tried making AI art, and there was definitely skill involved. I understand why prompt engineering can be hard. Every word I added could have unpredictable effects, constrain the space of images that can be generated because of the correlations. Choosing between similes was important.
What I realised was that I was fighting against the model to put my intent in the art. The model was extrapolating things I didn’t want to put in the art based on popularity and trends, and I was having to find workarounds to actually make what I wanted.
A pencil will never inject intent I don’t want into my art. My hand might slip, I might get the proportions wrong. But I’ll never try to draw a criminal and accidentally draw a black person because my pencil was exposed to more black criminals than white.
There are AI artists who can put their intent into their art. But they’re 1% of the people making AI art. Most of the time, a majority of the intention comes from a computer copying trends. And that’s not art. In order to make art, actual meaningful art with these programs, you have to be an expert.
Any idiot with a pencil is an artist. The pencil only does what you tell it. You don’t have to work to make it yours.
Any idiot with a pencil is an artist. The pencil only does what you tell it. You don’t have to work to make it yours.
I don’t like that we’ve settled here in defense of “human art” because that’s simply not true. No one, in the course of human history, has ever sat behind some kid doodling stupid pictures and clapped for them unless it was their mom. It’s a trope that real human artists are generally ignored through their lifetimes. Let’s stop pretending we’re all patrons of the arts and love everything that comes out of a person with a paint brush.
That said, you’re totally right, there’s 99% of people who don’t do shit with AI art, and don’t know what to do with it. There’s still an infinite capacity for a thing that makes whatever you can imagine to actually make something useful, I do truly believe that. If you haven’t seen “great AI art” it’s probably because there hasn’t been a great AI artist yet. Likening it back to photography, which was mostly a carnival trick and a thing for bored rich people, it took a long time before anyone ever identified a photographer who they thought actually showed them something that showed their subjectivity. We’re in the “old people staring at a camera looking bored” phase of this.
I’m not saying an idiot with a pencil will produce good art, or art that’s worth anyone else’s time. But they will always have the potential to draw something worth their own time. Anyone can engage in a worthwhile artistic endeavour using a pencil. It’s not about quality, or originality, or theming. It’s about intention. Any idiot can communicate what their own ideas with a pencil, on some level.
For example, take Adolf Hitler. He was a terrible artist. He had no understanding of proportion. His paintings don’t make logical sense. And you know what? That suits the man. Hitler was an idiot who couldn’t make logical sense of the world. His paintings reflect his terrible mind. There is that much value in them as an endeavour of self-expression. I can look at a painting by Hitler and say “that’s a Hitler”.
I can also look at an AI generated image by Highborn Flunky and say “that’s a Flunky”, but Flunky is a good artist. Hitler is not. That’s the difference I’m talking about. A painting reflects your soul even if you suck.
The difference between a camera and Stable Diffusion is that a camera makes it hard to convey artistic intent, while Stable Diffusion adds artistic intent that isn’t yours. Making art with a camera is hard, but anyone can know it when they see it. Making art with Stable Diffusion is an endeavour complicated by the fact that the AI is using mass manufactured intent to pass of intentless creations as art. If you fail to convey your intent with a prompt, you don’t get a dull scene, you get trendy bullshit.
And people are way more offended by being shown trendy bullshit by bad artists than boring scenes.
Yeah fair enough, I don’t disagree with you. I think for AI artists of the future it’ll be more about them asserting their subjectivity through the lens of each model (if that’s possible). Sort of in the same way that each camera has a character to it that is palpable across multiple people.
For example, take Adolf Hitler. He was a terrible artist. He had no understanding of proportion. His paintings don’t make logical sense. And you know what? That suits the man. Hitler was an idiot who couldn’t make logical sense of the world. His paintings reflect his terrible mind. There is that much value in them as an endeavour of self-expression. I can look at a painting by Hitler and say “that’s a Hitler”.
But I do disagree with this, though. Hitler may have been a terrible artist but he might be an equivalent artist in skill level to any number of other humans who might or might not have a “logical sense of the world”, “no understanding of proportion” etc. I don’t think Hitler’s art is good enough to even be able to tell anything about him other than that he doesn’t have an original voice yet.
I don’t think most (any?) AI artists have an original voice yet. Can they make one, even though the instrument makes most of (much of?) the music? I’m not bearish on the idea that humans can use this tool to make something useful I guess.
Oh, they can make something useful. I can watch Flunky make images with two very different checkpoints, and they have the same artistic style. And the breasts are enormous and heaving. It’s good for gooning to.
I just don’t believe we should have to put up with 99 average AI art users for every Highborn Flunky. I’m willing to give up gooning to Flunky pieces if it means I don’t have to see those awful sepia comics with the dot eyes on Lemmy.
And I’m concerned by men like Sam Altman and Elon Musk. I’m concerned by the energy cost of training, the economic costs to artists, and the effect on average people’s creativity. I would be willing to ban AI art in order to solve those issues. If 99% of humans can’t be responsible with the technology, then we shouldn’t have it. I feel the same way about automobiles, nuclear bombs, tanks, and plastic packaging.
Sure but it gets back to “Nobody wants to see anyone’s shit art”. No one wants to see anyone’s shit photos either. If it didn’t take like 5 minutes for a photo I bet most early photos would have been of tits, too. It doesn’t mean the tech can’t have a purpose and couldn’t be used in a flourishing, artistic way. It just means we need to do what we’ve always done and tell people with bad art to fuck off. :)
Humanity is too ignorant to recognise bad AI art on sight, every time. The trends are only easy to see in mass aggregate. That’s why individual action can’t stop it. We need aggregate action. We need to ban AI art.
I’ve tried making AI art, and there was definitely skill involved. I understand why prompt engineering can be hard. Every word I added could have unpredictable effects, constrain the space of images that can be generated because of the correlations. Choosing between similes was important.
What I realised was that I was fighting against the model to put my intent in the art. The model was extrapolating things I didn’t want to put in the art based on popularity and trends, and I was having to find workarounds to actually make what I wanted.
A pencil will never inject intent I don’t want into my art. My hand might slip, I might get the proportions wrong. But I’ll never try to draw a criminal and accidentally draw a black person because my pencil was exposed to more black criminals than white.
There are AI artists who can put their intent into their art. But they’re 1% of the people making AI art. Most of the time, a majority of the intention comes from a computer copying trends. And that’s not art. In order to make art, actual meaningful art with these programs, you have to be an expert.
Any idiot with a pencil is an artist. The pencil only does what you tell it. You don’t have to work to make it yours.
I don’t like that we’ve settled here in defense of “human art” because that’s simply not true. No one, in the course of human history, has ever sat behind some kid doodling stupid pictures and clapped for them unless it was their mom. It’s a trope that real human artists are generally ignored through their lifetimes. Let’s stop pretending we’re all patrons of the arts and love everything that comes out of a person with a paint brush.
That said, you’re totally right, there’s 99% of people who don’t do shit with AI art, and don’t know what to do with it. There’s still an infinite capacity for a thing that makes whatever you can imagine to actually make something useful, I do truly believe that. If you haven’t seen “great AI art” it’s probably because there hasn’t been a great AI artist yet. Likening it back to photography, which was mostly a carnival trick and a thing for bored rich people, it took a long time before anyone ever identified a photographer who they thought actually showed them something that showed their subjectivity. We’re in the “old people staring at a camera looking bored” phase of this.
I’m not saying an idiot with a pencil will produce good art, or art that’s worth anyone else’s time. But they will always have the potential to draw something worth their own time. Anyone can engage in a worthwhile artistic endeavour using a pencil. It’s not about quality, or originality, or theming. It’s about intention. Any idiot can communicate what their own ideas with a pencil, on some level.
For example, take Adolf Hitler. He was a terrible artist. He had no understanding of proportion. His paintings don’t make logical sense. And you know what? That suits the man. Hitler was an idiot who couldn’t make logical sense of the world. His paintings reflect his terrible mind. There is that much value in them as an endeavour of self-expression. I can look at a painting by Hitler and say “that’s a Hitler”.
I can also look at an AI generated image by Highborn Flunky and say “that’s a Flunky”, but Flunky is a good artist. Hitler is not. That’s the difference I’m talking about. A painting reflects your soul even if you suck.
The difference between a camera and Stable Diffusion is that a camera makes it hard to convey artistic intent, while Stable Diffusion adds artistic intent that isn’t yours. Making art with a camera is hard, but anyone can know it when they see it. Making art with Stable Diffusion is an endeavour complicated by the fact that the AI is using mass manufactured intent to pass of intentless creations as art. If you fail to convey your intent with a prompt, you don’t get a dull scene, you get trendy bullshit.
And people are way more offended by being shown trendy bullshit by bad artists than boring scenes.
Yeah fair enough, I don’t disagree with you. I think for AI artists of the future it’ll be more about them asserting their subjectivity through the lens of each model (if that’s possible). Sort of in the same way that each camera has a character to it that is palpable across multiple people.
But I do disagree with this, though. Hitler may have been a terrible artist but he might be an equivalent artist in skill level to any number of other humans who might or might not have a “logical sense of the world”, “no understanding of proportion” etc. I don’t think Hitler’s art is good enough to even be able to tell anything about him other than that he doesn’t have an original voice yet.
I don’t think most (any?) AI artists have an original voice yet. Can they make one, even though the instrument makes most of (much of?) the music? I’m not bearish on the idea that humans can use this tool to make something useful I guess.
Oh, they can make something useful. I can watch Flunky make images with two very different checkpoints, and they have the same artistic style. And the breasts are enormous and heaving. It’s good for gooning to.
I just don’t believe we should have to put up with 99 average AI art users for every Highborn Flunky. I’m willing to give up gooning to Flunky pieces if it means I don’t have to see those awful sepia comics with the dot eyes on Lemmy.
And I’m concerned by men like Sam Altman and Elon Musk. I’m concerned by the energy cost of training, the economic costs to artists, and the effect on average people’s creativity. I would be willing to ban AI art in order to solve those issues. If 99% of humans can’t be responsible with the technology, then we shouldn’t have it. I feel the same way about automobiles, nuclear bombs, tanks, and plastic packaging.
Sure but it gets back to “Nobody wants to see anyone’s shit art”. No one wants to see anyone’s shit photos either. If it didn’t take like 5 minutes for a photo I bet most early photos would have been of tits, too. It doesn’t mean the tech can’t have a purpose and couldn’t be used in a flourishing, artistic way. It just means we need to do what we’ve always done and tell people with bad art to fuck off. :)
Humanity is too ignorant to recognise bad AI art on sight, every time. The trends are only easy to see in mass aggregate. That’s why individual action can’t stop it. We need aggregate action. We need to ban AI art.