My view on AI art is neutral. It's a tool. Generally, without a learned skill in prompt design, you'll find the tool frustrating and give up. Without an understanding and appreciation for art outside the tool, you won't know where to go with the tool until you discover that appreciation. Without a knowledge of the artists who make the art that that AI tool uses, you will be hampered by your inadequacies in refining the artwork it produces.
As an artist, it is no more a threat to me than Photoshop or Illustrator, than Clip Studio Paint or Procreate, or than a Hunt 102 dip pen. AI tools for me are a rapid development aid and a quick way to test out ideas. A fail faster drill, if you will.
The people who fear AI tend to be the ones who are less proficient in their crafts, and therefore, are at a greater threat of displacement by those who choose to learn and develop their skills. They miss the point that human creativity drives all the works that are produced as "art", not the tools used.
That means the Hollywood producer who thinks he can have a press-button writer or artist is a fool who hasn't failed yet. "Yet" is the key. The untalented will never succeed by just "using AI."
The tool cannot create; it waits for the mind and hand of Man to drive it, exercising the creativity given to them by the Creator.
WM Briggs has a post too, that I had linked to in one of the Usury posts that had Little Shop of Horrors as the song.
Ultimately, if Large Language Models have no human inputs, they'll just learn from only the computer outputs, and go off the rails. They'll all be derivatives, and all be ugly as sin. They need the human, because only the human can determine and make the beautiful things to choose from for the inputs, and tell it what is beautiful in the outputs.
So, ultimately it needs the sacrifices of blood and work or it will spiral out of into bad art, bad writing, and bad music. And I'm fairly certain it would do so rather quickly.
I have to catch up on Briggs -- thanks for the tag. Missed last Thursday up through today. Life and everything.
As you said, AI is more susceptible to GIGO that almost anything else, because it has to take what it gets literally and act on it -- no reasoning or discernment. Human Creativity feeds AI Art, not the coders. They are just the mechanics making the tools. There are already anecdotal stories of C-suite Suits trying to get CS programmers to make art to-spec, with less than adequate results.
Another reason that Computing != Thinking & Feeling
I think the people who fear AI art and writing are the mediocre artists and writers.
As you say, AI will never replace a skilled human violinist - but it can replace "robotic" musicians the same way that it can replace a robotic sort of writer who is incapable of producing anything beyond glorified book reports.
Are you human? If so, great! Competing with a robot is easy, because there's really nothing of the divine spark within them that exists in human beings. The challenge, as it were, is for the robotic sort of artists, writers, musicians, etc. to really be human again.
Those who channel that divine spark into their work will stand out just as they always have.
100%. Second order effect: people will be less hesitant to check things out because there will be less retards polluting the pool. If the unoriginal robotic types do something else because they can’t compete with AI, there’ll be far fewer of them sending their work to people.
If publishers and all the rest are receiving less BS in their inbox, they’re less likely to dismiss out-of-hand anything worthwhile that shows up there just because they’re used to being flooded with BS.
Yes. Though I expect they'd have to pay for it. The stuff that would 'flood their inbox would be free stuff, produced by AI, which would be junk. Anything else will likely be paid.
I also see a jump in blue collar wages, as the AI wars turn into trade wars and trade shuts down overseas. But we'll see if that turns out to be the case.
Insofar as it goes, AI "art" puts together images from prompts based on what it has already scraped, which more and more consists of images that were generated from prompts, resulting in a convergence to stylistic sameness as it feeds on itself without new creative input from people.
I mostly object to AI art being called art - art requires choice. The role of the prompter isn't artist , it is patron, the interior designer picking which wallpaper, the company owner looking at logo options from the graphic designer and asking for "something like that but more funky". There is no actual human choice, and barring some randomness based off of what already exists, and the human input for promoting, no originality.
As I did in the piece, I would argue that this is the same saying that someone who's art solely involves taking pre-made pieces of art and modifying it isn't art and odd thing to say. That would mean a lot of the techno music isn't art, certainly those that mix it aren't, collaging isn't art, editing pre-made photography into a new picture isn't, etc.
The human making the prompts is a middle man, putting in the effort, and deciding the outcome. Whether he then goes on to edit it further or not determines if it's just an awful collage like a 3 y/o playing with scissors, or someone truly talented.
I think that this, right there, seems to be most people's issue - 99% of those they see playing with it are simply using the prompts to spit out unedited junk like I am. I get it, I'm putting out 3 y/o content. I'm a pickier 3 y/o, because I do have artistic training, but I'm also a father and husband, so I'm limited and don't have access to the editing software I got spoiled on, and the free stuff I've tried is simply so cumbersome in comparison I get frustrated.
But yes, it will all converge without human input, I completely agree. That's exactly the point Briggs made, and I'd concur. The machines need human blood and sacrifice, as I said on a previous post as well. Either the programmers will have to do it, or the artists, but likely both. Likely the programmers will actually have to tag things like "Rembrandt's paintings" to make sure that they only pull from him as a style if named and go through that kind of hassle. Or prompting will need to get more and more specific. Or both.
Thank you for your thought provoking piece. I disagree with your thesis, generally, and have some comments.
People can bicker over what ‘art’ is, as they can any definition. While on occasion important, arguments about definitions are still arguments about definitions - they distract from the point of the discussion.
However, I disagree with your analogy about techno. As someone who makes electronic music, it is far similar to composing music, than ‘taking and modifying’ but already I’m missing the point.
My main point is artistic integrity, rather than the generation itself and whether that is art. If you ask CHAT GPT to rephrase your writing, fine; that’s probably okay you can still lay claim to have created it yourself. What if, as will be possible soon, you get it to generate entire articles, using your voice; written exactly how you would? Can you claim you wrote that? Of course not. That would be creative fraud.
There is likely an arbitrary threshold, that if crossed, means you can no longer lay claim to be the primary creator of that piece. Putting in prompts, while skillful, is minor in the grand scheme of the artwork. As a result, to not deceive audiences, I believe one should be open about their AI use - but that’s just me.
The good news is that what I think will happen is the creation of some type of new artist, to keep up with the technology. Some form of super-prompter, or AI editor, that essentially, uses micromanaged prompts combining fields like music, writing, video, sound effects, to produce a film or tv show - or some other video/audio mixture.
That would be cool, and if the prompts are micromanaged and detailed enough, I’d happily call that art.
There is enough human input.
...
Hopefully I haven’t misrepresented your words - I cannot scroll up on my bloody phone to recheck the article. I wonder if Substack can introduce a ‘save comment’ feature.
JD Cowan has a post that tangents into this topic.
https://jdcowan.substack.com/p/whats-really-going-on
My view on AI art is neutral. It's a tool. Generally, without a learned skill in prompt design, you'll find the tool frustrating and give up. Without an understanding and appreciation for art outside the tool, you won't know where to go with the tool until you discover that appreciation. Without a knowledge of the artists who make the art that that AI tool uses, you will be hampered by your inadequacies in refining the artwork it produces.
As an artist, it is no more a threat to me than Photoshop or Illustrator, than Clip Studio Paint or Procreate, or than a Hunt 102 dip pen. AI tools for me are a rapid development aid and a quick way to test out ideas. A fail faster drill, if you will.
The people who fear AI tend to be the ones who are less proficient in their crafts, and therefore, are at a greater threat of displacement by those who choose to learn and develop their skills. They miss the point that human creativity drives all the works that are produced as "art", not the tools used.
That means the Hollywood producer who thinks he can have a press-button writer or artist is a fool who hasn't failed yet. "Yet" is the key. The untalented will never succeed by just "using AI."
The tool cannot create; it waits for the mind and hand of Man to drive it, exercising the creativity given to them by the Creator.
Thanks for the feedback. I'll check out the link
WM Briggs has a post too, that I had linked to in one of the Usury posts that had Little Shop of Horrors as the song.
Ultimately, if Large Language Models have no human inputs, they'll just learn from only the computer outputs, and go off the rails. They'll all be derivatives, and all be ugly as sin. They need the human, because only the human can determine and make the beautiful things to choose from for the inputs, and tell it what is beautiful in the outputs.
So, ultimately it needs the sacrifices of blood and work or it will spiral out of into bad art, bad writing, and bad music. And I'm fairly certain it would do so rather quickly.
I have to catch up on Briggs -- thanks for the tag. Missed last Thursday up through today. Life and everything.
As you said, AI is more susceptible to GIGO that almost anything else, because it has to take what it gets literally and act on it -- no reasoning or discernment. Human Creativity feeds AI Art, not the coders. They are just the mechanics making the tools. There are already anecdotal stories of C-suite Suits trying to get CS programmers to make art to-spec, with less than adequate results.
Another reason that Computing != Thinking & Feeling
The Briggs post was a ways back. He puts out so much content, I can only keep up with the ones that I really am interested in when I'm busy.
https://wmbriggs.substack.com/p/ai-needs-to-feed-on-fresh-meat-or
Ah, yes. I remember that one. Linked Roger Penrose's book in the comments.
I think the people who fear AI art and writing are the mediocre artists and writers.
As you say, AI will never replace a skilled human violinist - but it can replace "robotic" musicians the same way that it can replace a robotic sort of writer who is incapable of producing anything beyond glorified book reports.
Are you human? If so, great! Competing with a robot is easy, because there's really nothing of the divine spark within them that exists in human beings. The challenge, as it were, is for the robotic sort of artists, writers, musicians, etc. to really be human again.
Those who channel that divine spark into their work will stand out just as they always have.
Exactly. It clears out the dross, and makes it so that the field will be full of good work. People won't be able to make a living on BS stuff anymore.
100%. Second order effect: people will be less hesitant to check things out because there will be less retards polluting the pool. If the unoriginal robotic types do something else because they can’t compete with AI, there’ll be far fewer of them sending their work to people.
If publishers and all the rest are receiving less BS in their inbox, they’re less likely to dismiss out-of-hand anything worthwhile that shows up there just because they’re used to being flooded with BS.
Yes. Though I expect they'd have to pay for it. The stuff that would 'flood their inbox would be free stuff, produced by AI, which would be junk. Anything else will likely be paid.
I also see a jump in blue collar wages, as the AI wars turn into trade wars and trade shuts down overseas. But we'll see if that turns out to be the case.
Insofar as it goes, AI "art" puts together images from prompts based on what it has already scraped, which more and more consists of images that were generated from prompts, resulting in a convergence to stylistic sameness as it feeds on itself without new creative input from people.
I mostly object to AI art being called art - art requires choice. The role of the prompter isn't artist , it is patron, the interior designer picking which wallpaper, the company owner looking at logo options from the graphic designer and asking for "something like that but more funky". There is no actual human choice, and barring some randomness based off of what already exists, and the human input for promoting, no originality.
There is no craft.
As I did in the piece, I would argue that this is the same saying that someone who's art solely involves taking pre-made pieces of art and modifying it isn't art and odd thing to say. That would mean a lot of the techno music isn't art, certainly those that mix it aren't, collaging isn't art, editing pre-made photography into a new picture isn't, etc.
The human making the prompts is a middle man, putting in the effort, and deciding the outcome. Whether he then goes on to edit it further or not determines if it's just an awful collage like a 3 y/o playing with scissors, or someone truly talented.
I think that this, right there, seems to be most people's issue - 99% of those they see playing with it are simply using the prompts to spit out unedited junk like I am. I get it, I'm putting out 3 y/o content. I'm a pickier 3 y/o, because I do have artistic training, but I'm also a father and husband, so I'm limited and don't have access to the editing software I got spoiled on, and the free stuff I've tried is simply so cumbersome in comparison I get frustrated.
But yes, it will all converge without human input, I completely agree. That's exactly the point Briggs made, and I'd concur. The machines need human blood and sacrifice, as I said on a previous post as well. Either the programmers will have to do it, or the artists, but likely both. Likely the programmers will actually have to tag things like "Rembrandt's paintings" to make sure that they only pull from him as a style if named and go through that kind of hassle. Or prompting will need to get more and more specific. Or both.
Thank you for your thought provoking piece. I disagree with your thesis, generally, and have some comments.
People can bicker over what ‘art’ is, as they can any definition. While on occasion important, arguments about definitions are still arguments about definitions - they distract from the point of the discussion.
However, I disagree with your analogy about techno. As someone who makes electronic music, it is far similar to composing music, than ‘taking and modifying’ but already I’m missing the point.
My main point is artistic integrity, rather than the generation itself and whether that is art. If you ask CHAT GPT to rephrase your writing, fine; that’s probably okay you can still lay claim to have created it yourself. What if, as will be possible soon, you get it to generate entire articles, using your voice; written exactly how you would? Can you claim you wrote that? Of course not. That would be creative fraud.
There is likely an arbitrary threshold, that if crossed, means you can no longer lay claim to be the primary creator of that piece. Putting in prompts, while skillful, is minor in the grand scheme of the artwork. As a result, to not deceive audiences, I believe one should be open about their AI use - but that’s just me.
The good news is that what I think will happen is the creation of some type of new artist, to keep up with the technology. Some form of super-prompter, or AI editor, that essentially, uses micromanaged prompts combining fields like music, writing, video, sound effects, to produce a film or tv show - or some other video/audio mixture.
That would be cool, and if the prompts are micromanaged and detailed enough, I’d happily call that art.
There is enough human input.
...
Hopefully I haven’t misrepresented your words - I cannot scroll up on my bloody phone to recheck the article. I wonder if Substack can introduce a ‘save comment’ feature.
Anyway, thanks for the provoking piece.
The Delinquent Academic