For somebody who just wanted more good art, then they would welcome more tools for the sake of creativity and fast prototyping.
They do. But that's not what the generative ai slop is. The digital sculptors are largely all on board with ai powered auto uv unwrap and remeshing stuff for example. It's just tedious work that gets in the way of the more enjoyable creative parts of the process, is done algorithmically anyway by a lot of people, and depending on the workflow, doesn't have to be all that good, just passable. If you're making a 3d render of your collectable figurine, near enough is good enough for the UVs and mesh. You don't need it perfect like a video game or cgi movie face would need to be. And then when you go to 3d print it, the majority of people absolutely rely on some automations and algorithms for orientation, hollowing and supports. Because it's tedious. You'd go in and manually check it after, and maybe redo a few pieces, but yeah, people use the automated tools for this all the time.
They're generally on board (if they weren't grognard anti-digital purists) with ai for those purposes. We could call this the 'technical' or 'tool' or 'tedium' ai tools. The artists are generally on board with this style of ai, and more would be if it were only this. It's generative ai where the problems are.
Generative ai proponents say that it (generative ai) can be used as a tool to enhance or speed up creativity or prototypes or as a reference.
No it can't be used for references, because it's not accurate. AI hallucinates things. Anatomy and details. So do other artists of course, but for the artist seeking references, that's then a choice that make and you'd deliberately have some reference of both types, some of the references you'd grab are real photos, and then for the other references you'd grab other artist's work for style and potentially pose and composition or 'fun details' but with the understanding that the anatomy/details might be wrong or deliberately ignored in parts to make the pose and composition work, or just for the sake of it looking cool. And you might say that that's what AI does too. But it's not, it's creating a plausible guess based on text associations about what a thing might look like, without it actually being accurate. So it can't be used for the first use case, anatomy/detail reference. And for style and comp? Just use the artist's art. You typed their name into the prompt (or got around in-built restrictions on that with other tools), or used a lora that is trained only on their art. Why take the inferior copy when you've got their art right there, just use their art. So it's not great for the other use case either. The only place left for ai in terms of reference is a kind of paradoilic rorschach style of reference which some artists like to use, where the original idea is drawn out from splotches or randomness. I'm not usually a fan of this, it feels too 'fluffy/airy' for me, but I will occasionally draw ideas from the water on the glass in the shower for this, the steam and water coming together in shapes. Even if we grant that ai can help with that style of 'reference', for that you only need the general sense of it, and Dalle-1 the trial version that was shitting out indistinct nonsense could do this 4 years ago.
But not only is it effectively useless for reference, it's made searching for reference so much harder than it used to be. That's the other issue and part of why artists are getting so annoyed by it. Not only is it not much use, generative ai has actually made things worse. They want a good reference, and searching for that is just bringing up thousands of crap images. 'baby peacock' is the classic example at the moment, the one people are pointing to. Try googling it and you'll find a lot of garbage. But this is happening with everything and often in far subtler ways than the peacock thing making it actually even more annoying. Thousands of seemingly plausible but inaccurate images that we do not want hiding everywhere. Good and accurate reference is becoming harder to find because of the generative ai shitting up the internet (and google searching by date hasn't been accurate for a good 8 years or so). And so it's actually making art harder to do, right at the very start of the process.
It's a huge shame the right ceded the arts to the left, because this then creates a lack of understanding of the process on the right, and a desire to hurt them (artists), or at least an indifference to their struggles. And that's natural. But I would rather see good art reclaimed by the right, and the left's trash thrown out, rather than have both burnt down. Generative Ai is not helping make good art. It is hindering it.
(UVs are the squares that determine the surface of your 3d model, that you then paint or project the textures onto, once you’ve sculpted everything. Think of your model as an orange, and you want to unpeel it and make it flat with as few seams as possible and with those seams on the rear side or up at the naval of the orange, so there isn't a nasty seam in the colours. And you want all the squares generally even in shape and size so any textures aren't stretched out. It's just tedious to do)
Generative ai proponents say that it (generative ai) can be used as a tool to enhance or speed up creativity or prototypes or as a reference.
Because it is good for prototyping and concept art.
All your yapping following doesn't matter to that. Small inconsistencies are not important for that.
Not really AI, is it? More like an automated script that actually knows what you want doing.
Anatomy and details
AI got over the "six finger" thing about two years ago, my man. Provided you have a competent, trained model the errors you'll experience won't be that much worse than an average paid artist's.
The rest
I'm not going to address every point individually, just that most of the slop you see is the natural end result of granting internet access to hundreds of millions of turdworlders who lack the brain capacity to know what's good and what's slop, or even what's real and what's fake (as the three-headed African talent show singer has recently proven). You cannot grade what people with functioning neocortices desire over the basal lusts of those who are barely above animals.
The progression of AI as it develops is not going to be stopped. And the leaps and bounds it's made in the last 10 years is an incredible feat of engineering, given that in 2015 it could barely draw a cow in a field. And leftists are running scared because they thought they could colonize all of the arts and charge whatever they wanted for any old shit.
They do. But that's not what the generative ai slop is. The digital sculptors are largely all on board with ai powered auto uv unwrap and remeshing stuff for example. It's just tedious work that gets in the way of the more enjoyable creative parts of the process, is done algorithmically anyway by a lot of people, and depending on the workflow, doesn't have to be all that good, just passable. If you're making a 3d render of your collectable figurine, near enough is good enough for the UVs and mesh. You don't need it perfect like a video game or cgi movie face would need to be. And then when you go to 3d print it, the majority of people absolutely rely on some automations and algorithms for orientation, hollowing and supports. Because it's tedious. You'd go in and manually check it after, and maybe redo a few pieces, but yeah, people use the automated tools for this all the time.
They're generally on board (if they weren't grognard anti-digital purists) with ai for those purposes. We could call this the 'technical' or 'tool' or 'tedium' ai tools. The artists are generally on board with this style of ai, and more would be if it were only this. It's generative ai where the problems are.
Generative ai proponents say that it (generative ai) can be used as a tool to enhance or speed up creativity or prototypes or as a reference.
No it can't be used for references, because it's not accurate. AI hallucinates things. Anatomy and details. So do other artists of course, but for the artist seeking references, that's then a choice that make and you'd deliberately have some reference of both types, some of the references you'd grab are real photos, and then for the other references you'd grab other artist's work for style and potentially pose and composition or 'fun details' but with the understanding that the anatomy/details might be wrong or deliberately ignored in parts to make the pose and composition work, or just for the sake of it looking cool. And you might say that that's what AI does too. But it's not, it's creating a plausible guess based on text associations about what a thing might look like, without it actually being accurate. So it can't be used for the first use case, anatomy/detail reference. And for style and comp? Just use the artist's art. You typed their name into the prompt (or got around in-built restrictions on that with other tools), or used a lora that is trained only on their art. Why take the inferior copy when you've got their art right there, just use their art. So it's not great for the other use case either. The only place left for ai in terms of reference is a kind of paradoilic rorschach style of reference which some artists like to use, where the original idea is drawn out from splotches or randomness. I'm not usually a fan of this, it feels too 'fluffy/airy' for me, but I will occasionally draw ideas from the water on the glass in the shower for this, the steam and water coming together in shapes. Even if we grant that ai can help with that style of 'reference', for that you only need the general sense of it, and Dalle-1 the trial version that was shitting out indistinct nonsense could do this 4 years ago.
But not only is it effectively useless for reference, it's made searching for reference so much harder than it used to be. That's the other issue and part of why artists are getting so annoyed by it. Not only is it not much use, generative ai has actually made things worse. They want a good reference, and searching for that is just bringing up thousands of crap images. 'baby peacock' is the classic example at the moment, the one people are pointing to. Try googling it and you'll find a lot of garbage. But this is happening with everything and often in far subtler ways than the peacock thing making it actually even more annoying. Thousands of seemingly plausible but inaccurate images that we do not want hiding everywhere. Good and accurate reference is becoming harder to find because of the generative ai shitting up the internet (and google searching by date hasn't been accurate for a good 8 years or so). And so it's actually making art harder to do, right at the very start of the process.
It's a huge shame the right ceded the arts to the left, because this then creates a lack of understanding of the process on the right, and a desire to hurt them (artists), or at least an indifference to their struggles. And that's natural. But I would rather see good art reclaimed by the right, and the left's trash thrown out, rather than have both burnt down. Generative Ai is not helping make good art. It is hindering it.
(UVs are the squares that determine the surface of your 3d model, that you then paint or project the textures onto, once you’ve sculpted everything. Think of your model as an orange, and you want to unpeel it and make it flat with as few seams as possible and with those seams on the rear side or up at the naval of the orange, so there isn't a nasty seam in the colours. And you want all the squares generally even in shape and size so any textures aren't stretched out. It's just tedious to do)
Because it is good for prototyping and concept art.
All your yapping following doesn't matter to that. Small inconsistencies are not important for that.
All of that, and it would still be preferable to dealing with the internet "art" community.
They are just that obnoxious.
Cool story.
Not really AI, is it? More like an automated script that actually knows what you want doing.
AI got over the "six finger" thing about two years ago, my man. Provided you have a competent, trained model the errors you'll experience won't be that much worse than an average paid artist's.
I'm not going to address every point individually, just that most of the slop you see is the natural end result of granting internet access to hundreds of millions of turdworlders who lack the brain capacity to know what's good and what's slop, or even what's real and what's fake (as the three-headed African talent show singer has recently proven). You cannot grade what people with functioning neocortices desire over the basal lusts of those who are barely above animals.
The progression of AI as it develops is not going to be stopped. And the leaps and bounds it's made in the last 10 years is an incredible feat of engineering, given that in 2015 it could barely draw a cow in a field. And leftists are running scared because they thought they could colonize all of the arts and charge whatever they wanted for any old shit.