It's important to remember that there are many Monet paintings that critics don't like, or that aren't 'monet enough'. He painted fast to sell and make money and many think some paintings aren't as finished as they could be. He himself destroyed a number of water lily paintings before an exhibition [1], and again a lot of the work he did when he was partly blind due to cataracts.
It’s not a physical painting made by a well known artist.
It’s trying to hard to be a late Monet.
How much of our opinions are driven by context, rather than the actual subject? If Monet’s work is not so great without the context, is it still great? Or is context a critical piece of the art itself? Do we need to view a Monet piece within the scope of other Monet pieces, other artists, time periods, blindness, etc?
This feels like the example of (world-famous violinist) Joshua Bell playing violin in the DC subway and getting just a few bucks. It's totally different than paying money to see him in a concert hall, context matters so much...
I’d say for art, a lot? There’s a ton of art that a halfway decent painter could do now, the art of it was being the one to do it originally. At least that’s how I, as an absolute philistine in that regard, understand it ;)
I just saw a screenshot of someone pasting I think the US bill of rights to one of those AI write detectors; the site concluded that the text was written by AI.
There is a similar experiment where a famous violinist plays in a subway station. Nobody really notices or appreciates him and his music.
https://m.youtube.com/watch?v=hnOPu0_YWhw&ra=m
Yeah I agree, in art a lot is driven by context: there's so many paintings or songs that are not outstanding in itself, but the full human context around it makes it significant.
That brings up the idea that art can be "outstanding in itself", aesthetic in a vacuum, disconnected from what people are caring about. That's dubious, but anyway the AI art doesn't attempt that. Instead it has access to a lot of freeze-dried human context which it rehydrates and presents like a fresh meal, so it partially succeeds at providing that significance.
You're right. Maybe I should have said 'painting or songs that do not SEEM outstanding in itself". My point is that an AI 'rehydrating' human context that you mentioned, is (usually) not enough to get the same significance as human-made art.
For an edge case: people will be impressed and interested if you tell them that a piece was painted by an elephant, and then suddenly unimpressed if you tell them you were lying about that. So one function of art is as a sort of experiment, like the art is experimental data, where authenticity matters, because the interest is in the demonstration of a perspective, the reactions of an artist in the situation. Consider noir: a movie is much more plausibly authentic noir if it was made before about 1963, that is, if it was made by actors and directors who actually wore those hats (and lived through other tropes). Later on, it's imitation, regardless of how accurate: the experimental data is invalidated, it doesn't (seem to) mean so much.
I did a similar test back around 2007 on a famous photography website and shared real Edward Weston photographs of landscapes/buildings and people critiqued it quite negatively. This proved to me that popularity, context, and foreknowledge wins when it comes to art.
I think this HN commenter is also being fooled by the AI. It's likely that a lot of comments on HN are bots, so here you got an AI to comment about AI criticizing AI.
This is like asking people to rate this plate of bugs while serving them chicken. Even if tastes great, of course some people who will have a visceral reaction against it.
If we learn anything from all studies in this field, that is barely possible if not impossible at all, to change people’s mind. Even when they face clear evidence of their own mistake.
Not GP, but I think that's exactly the kind of bias that needs exposing. People are prone to holding a few experts/artists/objects/products in high regarding, defending/denying any flaws, while pushing down on those with less heritage.
I think it shows that art and how people relate to it is more complicated than you think. If the existence of a bunch of handpicked comments can lead you to your conclusion, then you will struggle to find any "good" art at all. Which may be an entirely coherent interpretation of the state of things; just not a very interesting one.
Okay, didn't mean to come out swinging, just a habit on internet forums anymore I guess.
To try to discuss more productively: I already thought Monet was overrated, and this was a long-held and considered opinion, not just a knee-jerk reaction to this post.
The post seemed to serve as a pepsi-challenge that confirmed what I already thought, but you're right that it's all cherrypicked anyway. That's the part I didn't consider so carefully.
That's fair. I on the other hand thought I was reading a dismissal based on nothing but some tweets (of all things), but in the end that was just an assumption I made.
I notice that were doing this underneath these words:
> Hopefully it helps people to at least be more conscious of their bias.
AI art enjoyers and missing the point of art: name a better duo.
No one has ever claimed AI cannot imitate a Monet, but however good the imitation, it still isn't art any more than a Xerox of a painting is art. This is the exact reason why most people feel bad after discovering that what they felt was work of human ingenuity, is just a fake, a simulacrum of it. The creation of art, arguably the most human of instincts, cannot be separated from the emotions and effort that went into it.
All this proves is that most people cannot tell if that picture is a Monet or not.
That's fair, but a lot of the attacks on AI art specifically point to perceived technical and compositional flaws. Heck, you still see people making "mangled fingers" jokes, and that hasn't been a thing in frontier models for a couple years now. Plus, a lot of the stylistic and "lacks creativity" critiques come from people churning out images with basic prompts on default settings; a modicum of effort makes it much more difficult to distinguish.
> All this proves is that most people cannot tell if that picture is a Monet or not.
It proves that people don't actually know what they like about "art" or even why they think some art is good, and some is bad.
These people criticized and trashed a widely regarded, famous painting because they were told that it was a cheap imitation.
If the AI generated a real imitation and the Met hung it on their walls I guarantee these same people would celebrate it just the same because they are told that it is real.
> All this proves is that most people cannot tell if that picture is a Monet or not.
It goes beyond that. It proves that many people have an inherent bias against AI itself that's unrelated to whatever it generates. "This was made by AI, therefore it's bad in every way".
... I mean, yes? People object to AI art (and generative AI in general) on ethical grounds, not just aesthetic ones. This is something anti-AI people are quite explicit about.
Good points, but consider what this post does prove: people’s arguments against AI art are shallow; they often attack the artifacts themselves instead of making your deeper argument.
I remember this old episode of Doctor Who where the Doctor scoffs at a postcard with the Mona Lisa on it and derides souless "art made by computers."
As a digital artist, of course I rolled my eyes at the time, but these days I just keep thinking about that storyline more and more.
We've basically transitioned to a world where digital art is almost the default, but I think the world is going to value physical art much more highly in the coming years.
Interesting how much the post sounds like an AI prompt itself. Are we all going to start talking like that? Think hard, make a plan, and only reply after deep consideration.
My guess is that the "experiment" isn't very rigorous -- just a direct link to some Twitter rando posting low-res image bait and cherry-picking answers (plus some trolly follow-ups).
Let this be an example of when you present your own work in real life. Context and framing is everything and does influence its interpretation and how people perceive your work. This has material effects on your life despite nothing objectively changing about the quality of your work.
Shows nothing about AI, shows a lot about how low the bar has fallen for not taking everything you see on social media at face value.
Enticing an easy and predictable knee jerk reaction from a couple dozen users also hardly proves anything.
Another sign that the context and the human factor will always play a huge role in how we experience art. For example, AI generated music can sound perfect, but still we value it less if we don't know anything about the musician's life.
It just proves serious bias and weakness of humans. We really "want" to believe we are special but in the end, you will soon never know if something is made by AI (autonomously or not) and it will trigger the same exact set of emotion as any human would, because in the end it changes nothing, we just want to believe it does (sadly).
Many artists have a distinct, original style. Originality is the ability to create novelty in a way which is aesthetically pleasing. I've yet to see AI create a single distinct style which is beautiful.
In my experience, an artist's original style arrives more or less from the artist having imitated and subsumed bits and pieces from other artists over the course of their artistic pursuit.
It isn't beautiful at all, nauseating uncanny valley stuff in my opinion, but AI images do have a style (or rather one of a number of idiosyncratic styles.) The sort of glossiness and unsettling focus in photorealistic images; the terrifying dreamlike surreality of more impressionist graphics; even the cartoony style used in corporate infographics. They're all quite distinctive to me at least and certainly aren't anything a human would produce. You can see the "influences" (i.e. stolen training data) but it really has come up with something itself.
Not that I think it should have. Kill it with fire and EMPs.
I know, but it could be AI-generated as well, because people can't tell them apart. The point was that even if AI could imitate Monet perfectly, it's not Monet. It's a worthless test.
Well, there's the Einstein test: can AI figure out general relativity if it's trained only on knowledge up to 1915 or so, before it was discovered. Similarly, you could do a Monet test: train AI on everything before Monet and try to get it to create paintings similar to Monet.
Original is something that is out of the data distribution. AI can't do anything original, because it's job is to imitate the data distribution.
Originality in itself is not hard, because pure noise is original. It should be original and beautiful.
The same thing in reverse:
https://i.kym-cdn.com/photos/images/newsfeed/003/240/678/2b9...
It's important to remember that there are many Monet paintings that critics don't like, or that aren't 'monet enough'. He painted fast to sell and make money and many think some paintings aren't as finished as they could be. He himself destroyed a number of water lily paintings before an exhibition [1], and again a lot of the work he did when he was partly blind due to cataracts.
[1] https://content.time.com/time/specials/packages/article/0,28...
1: the answers posted are cherrypicked to prove a specific point
2: some of the (albeit mislead) answers basically say "it's nice but it's not something a person willingly outlined and drew" and they are not wrong
3: some answers complain on the lack of depth and detail, color blurbs, and we have to agree the tested version is of very low resolution
so in the end we are left with: "some people who were told it was AI knee-jerked negatively" and i can't even start to see what's surprising about it
You're not surprised that people's judgement seems to be worthless?
the world is big, a given proportion of people will always behave strangely to the others
Two interesting replies:
It’s not a physical painting made by a well known artist.
It’s trying to hard to be a late Monet.
How much of our opinions are driven by context, rather than the actual subject? If Monet’s work is not so great without the context, is it still great? Or is context a critical piece of the art itself? Do we need to view a Monet piece within the scope of other Monet pieces, other artists, time periods, blindness, etc?
This feels like the example of (world-famous violinist) Joshua Bell playing violin in the DC subway and getting just a few bucks. It's totally different than paying money to see him in a concert hall, context matters so much...
> How much of our opinions are driven by context
I’d say for art, a lot? There’s a ton of art that a halfway decent painter could do now, the art of it was being the one to do it originally. At least that’s how I, as an absolute philistine in that regard, understand it ;)
I just saw a screenshot of someone pasting I think the US bill of rights to one of those AI write detectors; the site concluded that the text was written by AI.
There is a similar experiment where a famous violinist plays in a subway station. Nobody really notices or appreciates him and his music. https://m.youtube.com/watch?v=hnOPu0_YWhw&ra=m
Yeah I agree, in art a lot is driven by context: there's so many paintings or songs that are not outstanding in itself, but the full human context around it makes it significant.
That brings up the idea that art can be "outstanding in itself", aesthetic in a vacuum, disconnected from what people are caring about. That's dubious, but anyway the AI art doesn't attempt that. Instead it has access to a lot of freeze-dried human context which it rehydrates and presents like a fresh meal, so it partially succeeds at providing that significance.
You're right. Maybe I should have said 'painting or songs that do not SEEM outstanding in itself". My point is that an AI 'rehydrating' human context that you mentioned, is (usually) not enough to get the same significance as human-made art.
At least, for now.
For an edge case: people will be impressed and interested if you tell them that a piece was painted by an elephant, and then suddenly unimpressed if you tell them you were lying about that. So one function of art is as a sort of experiment, like the art is experimental data, where authenticity matters, because the interest is in the demonstration of a perspective, the reactions of an artist in the situation. Consider noir: a movie is much more plausibly authentic noir if it was made before about 1963, that is, if it was made by actors and directors who actually wore those hats (and lived through other tropes). Later on, it's imitation, regardless of how accurate: the experimental data is invalidated, it doesn't (seem to) mean so much.
I did a similar test back around 2007 on a famous photography website and shared real Edward Weston photographs of landscapes/buildings and people critiqued it quite negatively. This proved to me that popularity, context, and foreknowledge wins when it comes to art.
Seems the poster is the one fooled by the AI more than anything, because most likely the bulk of the replies are bots, so you got AI to criticize AI.
it was all engagement bait to auction off some NFT nonsense.
> Seems the poster is the one fooled by the AI
I think this HN commenter is also being fooled by the AI. It's likely that a lot of comments on HN are bots, so here you got an AI to comment about AI criticizing AI.
all right, all right, who got a bot to write this comment then?
bzzz, clank
This is like asking people to rate this plate of bugs while serving them chicken. Even if tastes great, of course some people who will have a visceral reaction against it.
But they’re confidently asserting a whole bunch of specific made up reasons this is shittier than a real Monet.
It’s like the sommeliers who can’t detect red vs. white wine when blindfolded.
People would come up with very specific made up reasons why they hate that plate of chicken as well, so I don't see your point.
As for your red vs. white wine comparison, it'd only make sense if one of those was doing its best to pretend to be the other one.
The point is the objections are bullshit. They can’t tell!
Then I guess you've completely missed my point.
If I hand you a lump of gold and tell you it's actually a piece of shit painted in gold, I'm sure you'll find reasons to hate it.
But if I make up reasons, that's bad!
"No thanks, I don't like shit" is fine.
"Oh yeah, I can definitely tell it's shit!" is not.
Very good. We need more of these experiments in all areas. Hopefully it helps people to at least be more conscious of their bias.
A proper experiment wouldn't have cherry picked the data. This seems more like rage bait than an experiment.
If we learn anything from all studies in this field, that is barely possible if not impossible at all, to change people’s mind. Even when they face clear evidence of their own mistake.
I do think this reveals peoples' biases, but not in the way you probably do.
I think Monet just wasn't as good as his renown purports.
EDIT: I doubt this experiment would go similarly for a Caravaggio or a Michelangelo.
Not GP, but I think that's exactly the kind of bias that needs exposing. People are prone to holding a few experts/artists/objects/products in high regarding, defending/denying any flaws, while pushing down on those with less heritage.
I think it shows that art and how people relate to it is more complicated than you think. If the existence of a bunch of handpicked comments can lead you to your conclusion, then you will struggle to find any "good" art at all. Which may be an entirely coherent interpretation of the state of things; just not a very interesting one.
It's a good thing then that I'm not concerned about how interesting my perspective is to you.
Of course. Sorry. Not your perspective, but the conclusion, if you leave it there. Which you well may.
Okay, didn't mean to come out swinging, just a habit on internet forums anymore I guess.
To try to discuss more productively: I already thought Monet was overrated, and this was a long-held and considered opinion, not just a knee-jerk reaction to this post.
The post seemed to serve as a pepsi-challenge that confirmed what I already thought, but you're right that it's all cherrypicked anyway. That's the part I didn't consider so carefully.
That's fair. I on the other hand thought I was reading a dismissal based on nothing but some tweets (of all things), but in the end that was just an assumption I made.
I notice that were doing this underneath these words:
> Hopefully it helps people to at least be more conscious of their bias.
Oh, well. Is this how it feels when it's working?
AI art enjoyers and missing the point of art: name a better duo.
No one has ever claimed AI cannot imitate a Monet, but however good the imitation, it still isn't art any more than a Xerox of a painting is art. This is the exact reason why most people feel bad after discovering that what they felt was work of human ingenuity, is just a fake, a simulacrum of it. The creation of art, arguably the most human of instincts, cannot be separated from the emotions and effort that went into it.
All this proves is that most people cannot tell if that picture is a Monet or not.
That's fair, but a lot of the attacks on AI art specifically point to perceived technical and compositional flaws. Heck, you still see people making "mangled fingers" jokes, and that hasn't been a thing in frontier models for a couple years now. Plus, a lot of the stylistic and "lacks creativity" critiques come from people churning out images with basic prompts on default settings; a modicum of effort makes it much more difficult to distinguish.
> All this proves is that most people cannot tell if that picture is a Monet or not.
It proves that people don't actually know what they like about "art" or even why they think some art is good, and some is bad.
These people criticized and trashed a widely regarded, famous painting because they were told that it was a cheap imitation.
If the AI generated a real imitation and the Met hung it on their walls I guarantee these same people would celebrate it just the same because they are told that it is real.
> It proves that people don't actually know what they like about "art" or even why they think some art is good, and some is bad.
That's because those are famously difficult questions to answer.
> All this proves is that most people cannot tell if that picture is a Monet or not.
It goes beyond that. It proves that many people have an inherent bias against AI itself that's unrelated to whatever it generates. "This was made by AI, therefore it's bad in every way".
that's because when dealing with art, the "why" something was made can be as important as the "what"
... I mean, yes? People object to AI art (and generative AI in general) on ethical grounds, not just aesthetic ones. This is something anti-AI people are quite explicit about.
Exactly. I'd rather appreciate a bad piece of art made in earnest by a person.
That's precisely the difference between art and a commodity.
Good points, but consider what this post does prove: people’s arguments against AI art are shallow; they often attack the artifacts themselves instead of making your deeper argument.
I remember this old episode of Doctor Who where the Doctor scoffs at a postcard with the Mona Lisa on it and derides souless "art made by computers."
As a digital artist, of course I rolled my eyes at the time, but these days I just keep thinking about that storyline more and more.
We've basically transitioned to a world where digital art is almost the default, but I think the world is going to value physical art much more highly in the coming years.
Interesting how much the post sounds like an AI prompt itself. Are we all going to start talking like that? Think hard, make a plan, and only reply after deep consideration.
https://xcancel.com/jediwolf/status/2054776716770320631
So odd this is flagged. Not complaining about moderation, but why would I bother submitting stimulating articles in future?
dang it would to use the flags as a way to prune recent HN users.
My guess is that the "experiment" isn't very rigorous -- just a direct link to some Twitter rando posting low-res image bait and cherry-picking answers (plus some trolly follow-ups).
NFTbro discovers expectancy effect. This has nothing to do with art or social experiments, so much so it's actually insulting to one's intelligence.
Let this be an example of when you present your own work in real life. Context and framing is everything and does influence its interpretation and how people perceive your work. This has material effects on your life despite nothing objectively changing about the quality of your work.
Shows nothing about AI, shows a lot about how low the bar has fallen for not taking everything you see on social media at face value. Enticing an easy and predictable knee jerk reaction from a couple dozen users also hardly proves anything.
“Made by Claude”
I think the more interesting thing going on here is the growing anti-AI sentiment. (Which I very much feel in myself too.)
That’s nothing new.
That’s just the art scene already ridiculed in the movie Interstate 60 with James Marsden and Gary Oldman and from 2002
https://youtu.be/HHwI37hkWfM?si=iFsWo3M5oSjLgE2F
Another sign that the context and the human factor will always play a huge role in how we experience art. For example, AI generated music can sound perfect, but still we value it less if we don't know anything about the musician's life.
It just proves serious bias and weakness of humans. We really "want" to believe we are special but in the end, you will soon never know if something is made by AI (autonomously or not) and it will trigger the same exact set of emotion as any human would, because in the end it changes nothing, we just want to believe it does (sadly).
Cherry picked, contrived, biased; in a word, slop.
Shows the pretentiousness of the twitterati more than anything else
Trading on pretentiousness in cliques has been a thing in art long before the internet and Twitter.
Being able to imitate Monet doesn't make you Monet. AI can't create anything original.
I loathe the blasted copyright washing machine as much as the next intellectually honest person, but:
> AI can't create anything original.
Can we? I mean, don't we base our output on experience and reprocess references + memories of things past to create what we deem as "new"?
Many artists have a distinct, original style. Originality is the ability to create novelty in a way which is aesthetically pleasing. I've yet to see AI create a single distinct style which is beautiful.
In my experience, an artist's original style arrives more or less from the artist having imitated and subsumed bits and pieces from other artists over the course of their artistic pursuit.
It isn't beautiful at all, nauseating uncanny valley stuff in my opinion, but AI images do have a style (or rather one of a number of idiosyncratic styles.) The sort of glossiness and unsettling focus in photorealistic images; the terrifying dreamlike surreality of more impressionist graphics; even the cartoony style used in corporate infographics. They're all quite distinctive to me at least and certainly aren't anything a human would produce. You can see the "influences" (i.e. stolen training data) but it really has come up with something itself.
Not that I think it should have. Kill it with fire and EMPs.
True. AI does have a style but it's more of a bug than a feature.
This is a real Monet.
I know, but it could be AI-generated as well, because people can't tell them apart. The point was that even if AI could imitate Monet perfectly, it's not Monet. It's a worthless test.
Define «original».
Under many definitions, where novel composition of existing knowledge or techniques is counted, it certainly can.
Well, there's the Einstein test: can AI figure out general relativity if it's trained only on knowledge up to 1915 or so, before it was discovered. Similarly, you could do a Monet test: train AI on everything before Monet and try to get it to create paintings similar to Monet.
Original is something that is out of the data distribution. AI can't do anything original, because it's job is to imitate the data distribution.
Originality in itself is not hard, because pure noise is original. It should be original and beautiful.
AI can absolutely create an original book, absurdly false.