Sorry, I wrote a response and deleted it. Let me try again.
I don’t know what exactly makes AI images so off-putting to me. The bare fact is that this image to me looks obviously AI-made and really unpleasant to see. I don’t know why some people react to AI images this way and others don’t.
My best guess is that AI images would begin to look more “cursed” to you if you spent some days or weeks drawing stuff with pencil and paper, maybe starting with some Betty Edwards exercises. But that’s just a guess, and maybe you’ve done that already.
I have some of the same feeling, but internally I’ve mostly pinned it to two prongs of repetition and ~status.
ChatGPT’s writing is increasingly disliked by those who recognize it. The prose is poor in various ways, but I’ve certainly read worse and not been so off-put. Nor am I as off-put when I first use a new model, but then I increasingly notice its flaws over the next few weeks. The main aspect is that the generated prose is repetitive across the writings which ensures we can pick up on the pattern. Such as making it easy to predict flaws.
Just as I avoid many generic power fantasy fiction as much of it is very predictable in how it will fall short even though many are still positive value if I didn’t have other things to do with my time.
So, I think a substantial part is that of recognizing the style, there being flaws you’ve seen in many images in the past, and then regardless of whether this specific actual image is that problematic, the mind associates it with negative instances and also being overly predictable.
Status-wise this is not entirely in a negative status game sense. A generated image is a sign that it was probably not that much effort for the person making it, and the mind has learned to associate art with effort + status to a degree, even if indirect effort + status by the original artist the article is referencing.
And so it is easy to learn a negative feeling towards these, which attaches itself to the noticeable shared repetition/tone. Just like some people dislike pop in part due to status considerations like being made by celebrities or countersignaling of not wanting to go for the most popular thing, and then that feeds into an actual dislike for that style of musical art.
But this activates too easily, a misfiring set of instincts, so I’ve deliberately tamped it down on myself; because I realized that there are plenty of images which five years ago I would have been simply impressed and find them visually appealing. I think this is an instinct that is to a degree real (generated images can be poorly made), while also feeding on itself that makes it disconnected from past preferences.
I don’t think that the poorly made images should notably influence my enjoyment of better quality images, even if there is a shared noticeable core. So that’s my suggestion.
‘Repetition’ is certainly a drawback to the ChatGPT style: we have lost em dashes and tricolons for a generation. But it can’t in its own right explain the reaction to the SD image, because… ‘German Expressionist linocut’ just doesn’t describe a default, or even a common, output style of any image generative model ever. (That’s part of why I like to use ‘linocut’ as a keyword, and for better or worse, people who might reach for ‘German Expressionist’ these days typically reach for Corporate Memphis instead.)
It could however be a kneejerk reaction: “oh no, this is a generated image, therefore it is exhaustingly overused and boring [even if it isn’t actually]”.
Sorry, I wrote a response and deleted it. Let me try again.
I don’t know what exactly makes AI images so off-putting to me. The bare fact is that this image to me looks obviously AI-made and really unpleasant to see. I don’t know why some people react to AI images this way and others don’t.
My best guess is that AI images would begin to look more “cursed” to you if you spent some days or weeks drawing stuff with pencil and paper, maybe starting with some Betty Edwards exercises. But that’s just a guess, and maybe you’ve done that already.
I have some of the same feeling, but internally I’ve mostly pinned it to two prongs of repetition and ~status.
ChatGPT’s writing is increasingly disliked by those who recognize it. The prose is poor in various ways, but I’ve certainly read worse and not been so off-put. Nor am I as off-put when I first use a new model, but then I increasingly notice its flaws over the next few weeks. The main aspect is that the generated prose is repetitive across the writings which ensures we can pick up on the pattern. Such as making it easy to predict flaws. Just as I avoid many generic power fantasy fiction as much of it is very predictable in how it will fall short even though many are still positive value if I didn’t have other things to do with my time.
So, I think a substantial part is that of recognizing the style, there being flaws you’ve seen in many images in the past, and then regardless of whether this specific actual image is that problematic, the mind associates it with negative instances and also being overly predictable.
Status-wise this is not entirely in a negative status game sense. A generated image is a sign that it was probably not that much effort for the person making it, and the mind has learned to associate art with effort + status to a degree, even if indirect effort + status by the original artist the article is referencing. And so it is easy to learn a negative feeling towards these, which attaches itself to the noticeable shared repetition/tone. Just like some people dislike pop in part due to status considerations like being made by celebrities or countersignaling of not wanting to go for the most popular thing, and then that feeds into an actual dislike for that style of musical art.
But this activates too easily, a misfiring set of instincts, so I’ve deliberately tamped it down on myself; because I realized that there are plenty of images which five years ago I would have been simply impressed and find them visually appealing. I think this is an instinct that is to a degree real (generated images can be poorly made), while also feeding on itself that makes it disconnected from past preferences. I don’t think that the poorly made images should notably influence my enjoyment of better quality images, even if there is a shared noticeable core. So that’s my suggestion.
‘Repetition’ is certainly a drawback to the ChatGPT style: we have lost em dashes and tricolons for a generation. But it can’t in its own right explain the reaction to the SD image, because… ‘German Expressionist linocut’ just doesn’t describe a default, or even a common, output style of any image generative model ever. (That’s part of why I like to use ‘linocut’ as a keyword, and for better or worse, people who might reach for ‘German Expressionist’ these days typically reach for Corporate Memphis instead.)
It could however be a kneejerk reaction: “oh no, this is a generated image, therefore it is exhaustingly overused and boring [even if it isn’t actually]”.