AI Art Is the New Stock Image(ia.net)
ia.net
AI Art Is the New Stock Image
https://ia.net/topics/ai-art-is-the-new-stock-image
14 comments
Right so the author has taken the most extreme level of weirdness and tried to use it as an example of normal image generator output. Thats quite dishonest.
Totally anecdotal, just tonight i saw a video my child was watching with generated images. At first i thought, I've seen that face in generated images, i think, then we looked a little closer, and noticed obvious mistakes (extra or weird fingers, weird body positions, cross eyes where you wouldn't expect them).
Of course it's the first time I've spotted it in the wild where it wasn't explicitly marked as generated content, so maybe i've missed a bunch.
Of course it's the first time I've spotted it in the wild where it wasn't explicitly marked as generated content, so maybe i've missed a bunch.
That totally happens, and its not like the blog post is definitely wrong. Its just dishonest, which undermines its own argument.
"One of the lamest ways to use AI is to generate images in the style of painting artists."
The lamest way? It is my favourite! I guess im a lamer then. :(
The lamest way? It is my favourite! I guess im a lamer then. :(
This reminds me of when people claim they don’t understand why people get plastic surgery because it makes people look so bad.
I literally don't understand why people get plastic surgery because it makes people look so bad! So can you explain?
I did it, got pretty much my entire face done, skin and bone. The only part untouched was the skin above my upper lip. I went from a 4 on a good day to a 7 but probs more because I'm my own worst critic. You would never know I had anything done unless you knew from the perfectly ideal nose curve. 6 hours on the operating table, 4 days of misery, couple of metal plates, some screws, and 6 months ish of not being able to feel the top of my head which was more novel than anything else. Best money I've ever spent.
A lot of the photos you see of "plastic surgery face" are before the swelling comes down and you look normal again.
A lot of the photos you see of "plastic surgery face" are before the swelling comes down and you look normal again.
There are plenty of examples of younger people with obvious bad results. The one going around on social media right now is Erin Moriarty who plays Starlight on the Amazon show The Boys. The issue is that some combination of too many procedures, poor choice of procedures, bad luck, and bad surgeons can result in an obvious uncanny valley effect. Once someone winds up in uncanny valley territory, it seems near-impossible to correct.
https://knowyourmeme.com/memes/erin-moriarty-plastic-surgery
https://knowyourmeme.com/memes/erin-moriarty-plastic-surgery
Same reason you see the bad CGI in movies. When it's good, you don't.
People that get high quality facial work, especially when they are younger, and don’t have negative reactions to don’t stand out as “people that had plastic surgery”.
At a certain point it catches up to you, but it’s very feasible for a 40 year old to look 30.
At a certain point it catches up to you, but it’s very feasible for a 40 year old to look 30.
The usual wisdom here is that people often see what they think is X_1 post application of plastic surgery and ask why they they'd do that to the beautiful X_0 they were. The reality is often that people are talking about how X_35 is so much worse than X_34 and how X_34 was the epitome of natural beauty.
So it's a stopping problem. One might as well ask "You bought your Bitcoin at $0.01?! Why didn't you sell it when it was $60k?!".
So it's a stopping problem. One might as well ask "You bought your Bitcoin at $0.01?! Why didn't you sell it when it was $60k?!".
Good taste, or lack thereof: https://xkcd.com/915/
It's possible to become a "connoisseur" at anything. Music, art, movies, food, wine, or physical attractiveness.
People with poor taste[1] get bad plastic surgery, and are rewarded for these decisions because members of the opposite sex with similarly poor taste become more attracted to them.
This is the equivalent to junk food being loaded up with salt, fat, and sugar: it works on a surprisingly large fraction of the population.
Personally, I much prefer a fancy meal that involves spices, herbs, and visual appeal, but I'm actually in the minority in this respect.
Similarly, I prefer elegantly dressed women with air of grace about them, but that's just me and my refined taste. There are enough men out there that simply want bigger boobs, that this has spawned a multi-billion-dollar plastic surgery industry. Not to mention push-up bras, etc...
[1] Some people would argue that all taste is relative, that there is no right or wrong here, etc... Normally I would agree, but we've all seen the "People of Wallmart", and... well... yeah.
It's possible to become a "connoisseur" at anything. Music, art, movies, food, wine, or physical attractiveness.
People with poor taste[1] get bad plastic surgery, and are rewarded for these decisions because members of the opposite sex with similarly poor taste become more attracted to them.
This is the equivalent to junk food being loaded up with salt, fat, and sugar: it works on a surprisingly large fraction of the population.
Personally, I much prefer a fancy meal that involves spices, herbs, and visual appeal, but I'm actually in the minority in this respect.
Similarly, I prefer elegantly dressed women with air of grace about them, but that's just me and my refined taste. There are enough men out there that simply want bigger boobs, that this has spawned a multi-billion-dollar plastic surgery industry. Not to mention push-up bras, etc...
[1] Some people would argue that all taste is relative, that there is no right or wrong here, etc... Normally I would agree, but we've all seen the "People of Wallmart", and... well... yeah.
I don’t know what your angle is but if plastic surgery is done right (ie most cases) you never explicitly notice it.
if you had before and after photos it would certainly be noticable. Otherwise what would be the point? You cant tell me botox, lip injections or a nose job are "not explicitly noticable" when they drastically change your facial profile, especially over multiple years
Yes, but I’m talking about in the wild which is what is relevant to this analogy.
it's not gonna stop people being smug and dismissive of ai images any time soon. their loss though
The article lack substance, it doesn't offer a substantive critique and the cherry picked examples are both confusing and arbitrary. I say this as someone deeply critical of the way deep learning models are being trained and deployed, and deeply fearful of how they'll distort our culture in the future. But we need substantive engagement with the aesthetic and economic issues. Such a shallow critique has no place on hackernews.
Yes, it's the new stock image, but one thing I think is a game changer is the incorporation of custom text. I am a member of a homeschool group, and they use impressive AI-generated stock images that always have the group name or other relevant text embedded in creative ways, like on a chalkboard or whiteboard.
I don’t get the negativity in these comments. Sure, this is not a groundbreaking article, just a simple blog post with a simple plea, stated very clearly in the conclusion:
> AI or not, images should be given the same attention as text. Choose each visual carefully and place it in the right context. Don’t let AI do your job, use it as a tool.
I too have seen a proliferation of AI art accompanying articles around the web and many of them, as this article points out, are poorly selected and bad replacements of the already bad practice of every single article being accompanied by a photo. Likely for SEO or something.
> AI or not, images should be given the same attention as text. Choose each visual carefully and place it in the right context. Don’t let AI do your job, use it as a tool.
I too have seen a proliferation of AI art accompanying articles around the web and many of them, as this article points out, are poorly selected and bad replacements of the already bad practice of every single article being accompanied by a photo. Likely for SEO or something.
It it's produced by AI it's not art by definition.
The legal world already has an answer to machine generated art as it pertains to copyright and it somewhat agrees with you. The act of a machine generating images isn't copyrightable but the act of selecting one or more of those images becomes copyrightable art.
I've still not seen a convincing answer from the legal world as to how MUCH human involvement makes an AI-assisted image copyrightable. Can I claim copyright on my picture of a pelican in a fancy hat or not?
https://fedi.simonwillison.net/@simon/111489351875265358
https://fedi.simonwillison.net/@simon/111489351875265358
Which definition? I thought art was notoriously difficult to clearly define!
The Oxford dictionary defines art: ” the expression or application of human creative skill and imagination, typically in a visual form such as painting or sculpture, producing works to be appreciated primarily for their beauty or emotional power.”
Having spent some passionate time with StableDiffusion I can outline a few signs in the details of images this article has missed.
- tiny details never make sense. In real art or photos, no line or edge is useless or senseless. But since generated images are only hallucinations, the tiny lines often blend into something else, convert to something else, or simply fade out.
- scaling is wrong. Most of the times scaling won't make sense. Take a tiny part of the image, just looking at that part, can you guess how big or small other things in that image should be? A box lying on the floor looks way too small or big for example.
- perspective is off. That sponge + cheese image for example the angle of eyes or head or the appearant thickness on that angle isn't right.
Models are getting really good though. Previously you could look at texture of clothes and tell, that is almost gone these and you have to find some other pattern which is usually some other non-symmetry.
Tiny details, textures is the most visible and almost always recognisable sign of a generated (unaltered) image.
- tiny details never make sense. In real art or photos, no line or edge is useless or senseless. But since generated images are only hallucinations, the tiny lines often blend into something else, convert to something else, or simply fade out.
- scaling is wrong. Most of the times scaling won't make sense. Take a tiny part of the image, just looking at that part, can you guess how big or small other things in that image should be? A box lying on the floor looks way too small or big for example.
- perspective is off. That sponge + cheese image for example the angle of eyes or head or the appearant thickness on that angle isn't right.
Models are getting really good though. Previously you could look at texture of clothes and tell, that is almost gone these and you have to find some other pattern which is usually some other non-symmetry.
Tiny details, textures is the most visible and almost always recognisable sign of a generated (unaltered) image.
Tiny details, textures is the most visible and almost always recognisable sign of a generated (unaltered) image.
In some ways, I think AI art resembles what you'd see in a dream. The "big picture" is there and logical, but otherwise it's full of little inconsistencies that don't make sense.
In some ways, I think AI art resembles what you'd see in a dream. The "big picture" is there and logical, but otherwise it's full of little inconsistencies that don't make sense.
> In real art or photos, no line or edge is useless or senseless.
In real art, artists screw up little details all the time. I've recently seen an X thread (unfortunately no longer have the link) where the author quite convisiningly highlighted all the kinds of typical AI artefacts found in the works of Rembrandt.
In real art, artists screw up little details all the time. I've recently seen an X thread (unfortunately no longer have the link) where the author quite convisiningly highlighted all the kinds of typical AI artefacts found in the works of Rembrandt.
I am not an artist but can draw a little and if by screwing up little details you mean filling the image with random strokes which make the bigure picture look correct than those are not what I am talking about. In generated image when you follow a tiny detail it will just disappear into something else. In painting you can usually see the strokes up close but in a gen image, those strokes will be a weird blend if they are visible. Tiny details like paint smidges etc won't be there. Shadows of tiny details won't make sense.
I really like to see that X thread.
I really like to see that X thread.
I only tested the pulp example, but it was trivial for me to create images that look much closer to pulp covers than the provided image. The author of this blog clearly has zero experience with image generators.
Even before Stable Diffusion et al, I had grown to detest the majority of stock image use (for articles/blogs). They almost never add to the article, and often mislead.
It's not a coincidence that the one place they are heavily used is marketing/ads. When people add a stock image to an article/blog post, they really are playing the ad game, even if they're not selling for money.
Not a big loss if they're all replaced with AI.
It's not a coincidence that the one place they are heavily used is marketing/ads. When people add a stock image to an article/blog post, they really are playing the ad game, even if they're not selling for money.
Not a big loss if they're all replaced with AI.
I wish articles like this would provide some indication of who wrote them. This is an opinion piece, so I'd like to know who it is the opinion of!
Personally, I find that the most satisfying AI images are not the ones that try to imitate photographs or do a hyperrealistic style, but the ones inspired by artistic schools which were intentionally coarse and textured.
A few examples here, from back when I was a Midjourney user:
https://x.com/felipeerias/status/1749270656041234656
A few examples here, from back when I was a Midjourney user:
https://x.com/felipeerias/status/1749270656041234656
Some examples:
Pepsi Buffalo Wings flavor: https://www.reddit.com/r/weirddalle/comments/198g30z/which_o...
Razer gaming product concepts: https://www.reddit.com/r/weirddalle/comments/197qgi1/with_th...
Indiana Jones stealing subs: https://www.reddit.com/r/weirddalle/comments/1987cj4/running...