AI is creating a 'human premium' for art created by people | Hacker News Reader

AI is creating a 'human premium' for art created by people

bbc.com

74 pointscolinprince122 comments

Art was never separate from the person and the culture that created it. It was its soul. Now our lives are flooded with soulless creations night and day--the product of lazy minds and an amalgamation of stolen already-bland cultural works.

So just like before AI.

And now it has been exacerbated further

This is such a lazy "gotcha" argument, and a prevalent one at that. Some people seem to think that anyone who is upset about the massive deluge of garbage AI has enabled was somehow perfectly happy with the previous status quo. "Things were already bad so who cares if we make them worse?" is a sentiment that seems does have some wider acceptance in the software world, but not everyone feels this way.

No it is not lazy. But if you devoured the marvel slop or gray's anathomy in which production no human intelligence was utilized, what difference does it make if the slop is AI made?

If you filled your day with brainrot slop - does it matter the origin. Does it matter if Instagram thirst traps are so with surgery and filters or if the whole meat puppet stuff is skipped

the difference is it’s worse

> If you filled your day with brainrot slop - does it matter the origin. Does it matter if Instagram thirst traps are so with surgery and filters or if the whole meat puppet stuff is skipped

So far it seems like AI video generation will vastly multiply the quantity of crap like this and enable it to be even worse while being cheaper and easier to produce, while not meaningfully improving the status quo for higher quality works.

"Does it matter which you watch?" Well, who said that was the most interesting question in the first place? Who makes slop is not the only way the situation is changing.

What does it mean for art to have a soul?

I would argue that modern art has none. Further, I would argue that art is what speaks to people.

That means that an AI generated image which makes people laugh has more soul than whatever the top selling modern artists was in 1995. By my estimation, that makes a lolcat a better piece of art than most modern art.

That is something that I actually think is true. Of course there are great artists who can show us much, much more than what is inside a lolcat. Its just that I haven't seen one of those in a while.

Artists are concerned for their jobs; That I get. Trying to defend art by claiming magical things like it doesn't have any soul is just cope. And frankly, the more I see of that, the less I care about artists and their real problems.

The soul is whatever story the buyer puts into it. A few years ago, people attributed soul to NFTs, but I think in the long run, humans will ultimately be able to generate more pervasive stories. There are thousands of artists in towns continuing to generate them.

NFTs were just a supposed new way to claim ownership over works of art. NFTs themselves were not a new artform in itself. I doubt anyone claimed that NFTs had some kind of artistic value, just by virtue of being an NFT.

Modern art refers to art produced roughly from the 1860s to the 1970s.

people who complain about modern art almost always know essentially nothing about art. Usually heavy overlap with those that complain about anything that is different from what they like/grew up with

Because the public never gets the opportunity to be exposed to good art, whether classic or abstract or whatever style. Most what's called "modern art" in the style of early 20th century like Picasso or later Pollock is low quality sewage, but that's what's being fed to the public. That's what everybody grew up with and everybody hates.

The public is underexposed to beautiful and interesting art, but who is to blame?

Yes and yet modern and contemporary are synonyms. See polysemy. The context makes it clear.

You're confusing entertainment with art.

Art is a form of expression and/or communication. It's interesting because it shows us different perspectives, as well as things that overlap with ours. When you see a smiley face on your lawn and think it's a sign by the person you just fell in love with, and then find out the exact same artifact was in fact created by wind sweeping some branches together, you'll feel different about it. Similar for art. Or take jokes. A friend might find a joke of yours funny they might not find funny from a person they dislike. That's not "unfair" or "objectively wrong", just like rejecting the same artifact depending on source. If you don't understand this, you talking about "magic" is the cope.

> Its just that I haven't seen one of those in a while.

So that's the argument for making sure of that? Heh.

> Artists are concerned for their jobs; That I get.

The article isn't about artists being concerned for their jobs. This is about people rejecting slop.

To me it means more that art is a window into its creator. I am more interested in exploring how and why a piece of art is being communicated, why certain details were used, how constraints were incorporated. AI is usually boring to me. By generating so much peripheral and extreme detail it removes rather than enhances expression in many cases. I think that's what people mean by missing a soul. It becomes difficult to tell what was intentional expression or a choice of the artist.

For 3 years now, I've gotten into arguments with both AI bros and artists about this. The main point of art is its cultural space (its creator, the zeitgeist etc.).

A human premium, and an elevation of human made art was always inevitable. Most "AI art" token spend is TPS reports, and explainer slides, maybe educational materials on the fringes. Maybe there is a chance some forms of AI art becomes its own new medium (like theater vs "the talkies") but it was never going to be a replacement.

Just 3 years of pointless hand-wringing, headlines and breathless interviews for nothing.

"Aaron Sorkin says he will never use AI".

Yes. No one asked him to! But alas, clickbait rules all.

I recently had an idea for procedural music that would be a fusion of classical ideas of Process ala Glass/Reich/Hisaishi with music making techniques and possibly the interactivity of video games. No idea if it'll work it was more a shower thought than a deep creative endeavour that I'm attached to but my point is that art is art and AI will just be a tool in the end for true artists.

This kind of snobbery sucks. You give the same reason for art created by humans without interacting with AI being unreachable as you do for AI created art being bad. Human made art is "not separate from culture" while AI made art is "amalgamation of already bland cultural works".

Overall, the perspective espoused here is bad for art. Art isn't hidden inside effort, you don't find it by saying "don't be lazy", and if it's inseparable from where it came from (which I think is pretty debatable, art travels through time and the world and can find interest), this is as true for what humans make as what AI makes.

I think the last bit gives the game away. This perspective is an otherwise content free way of saying "some art is interesting to me, and some isn't".

It’s not snobbery. Art is just another way for humans to talk with each other, if AI does all the talking, people are largely not interested.

>saying "some art is interesting to me, and some isn't".

That applies to every single human being, including you, except maybe for art scholars.

That last part is what I just said. The original post is saying nothing aside from something that's already true of everyone, and has nothing to do with whether AI can or does make interesting things.

I'm also not sure that art is actually a way that humans talk to each other. That assumes that the point is delivering some information, but a lot of artists talk about it as a compulsion, or something they do for themselves, or as being about understanding the world. I would actually say that many would say that it's definitively not about audience reception. Simultaneously, audiences interpret things of their own accord all the time. Unlike when sending a message, art is often most fully interacted with when you're engaging with things that weren't put there explicitly.

You assume that humans exchange information only consciously. Someone can feel a compulsion to capture the beauty of something, and someone can still read it as “wow, so this is what beauty can be like”. Or they can read something else, human comms are imperfect and that’s part of it.

What do you see in AI art that you love or find interesting so much?

Also, while I disagree about your point on the audience part — if that is your belief, the article and my or anyone else’s opinion should be absolutely irrelevant for artists. They feel compulsion, the do their AI art. Why does it matter if it finds its audience or sells? Make art, put your soul into it, put your LLM’s soul into it if you believe it has it, make it for all the interesting and cool things. Show it to humans, show it to bots, don’t show it and keep it on your hard drive — the audience doesn’t matter if AI/you made something interesting and artistic. So what if someone wants a “human-made” sticker on whatever drawing, an artist’s goal is to make art, no? Not get anyone’s approval. Why does someone’s “snobbery” matter in this regard at all?

>That last part is what I just said. The original post is saying nothing aside from something that's already true of everyone, and has nothing to do with whether AI can or does make interesting things.

Well, your post, following your own logic, just says that you like AI-generated stuff. You cannot judge what's good or bad for art, you can't call someone preference "snobbery" because the only property of art is that someone likes it. Interesting things is also then completely subjective.

There are several cases of people saying they hate gays and being caught in a gay brothel, people saying they never watched porn and caught on camera with their phone open to trans porn, etc.

It works both ways. There are plenty of people who fake orgasm. :)

This kind of philistinism sucks.

Arguing that the artist's own hardwork, and through that their personality and soul, is integral to art is not snobbery: snobbery is stating that "ordinary" art is beneath you.

You didn't make that argument at all. You said our lives are flooded with soulless creations night and day--the product of lazy minds. You called people you dont like lazy.

You insult people and then lie about what you said.

If I read an interesting story, it might lead to some sort of moral or dramatic tension or experience that sticks with me. It could be years later when I live some particular experience and the author's intent becomes clear, or I get some insight into why they had a certain character act in a certain way, or maybe that I missed the point entirely and I am seeing the subtly now.

All of which are impossible with AI. It has no intent, it fulfulls the ask from the prompt. If the prompt is so detailed to imbue the work with deeper meanings, then the author is the person that wrote the prompt, not the LLM which becomes something of a translater.

If LLMs get so good at architecting stories that they can whip up some conceptual layer of understand that feels like communication then we will have achieved a higher form of literary remixing. LLMs are still a series of weights,they still operate in a stateless way and have no initiative besides our prompt. Unless we change that, the value of what an LLM can produce remains limited.

To think otherwise is like a tribesman finding a radio and thinking they are hearing the voice of God. Because you aren't following where the words come from you invent a persona for something that is fundamentally a vehicle of transport.

All of this line of reasoning in the second half is predicated on the idea that there's nothing besides the prompt, but that's an architectural decision that labs optimize for, not something inherent to the field or methods. It's actually fairly trivial to allow for unprompted or accumulated work, except that many people then get mad at it because "I didn't ask for it to do XYZ!"

Your insight years later is still something you constructed, it's a new story that you've told yourself, presumably without any new input from the writer, all of which you can do with any given piece of media. You can do that without the intent having been there on the writer's part at all, even, something which is fairly common to do.

So you think art is overwhelmingly for interpretation. I'm trying to put your comments together and get a handle on this viewpoint. You said elsewhere that artists aren't in it to communicate to the audience. So the artist uses communication media and tropes to do something cryptic and personal, which everybody then misunderstands but enjoys using as a vehicle for their own, how did put it, "story that you've told yourself" about an imagined artist. With the result that a completely phony artificial* artist will suffice. Right?

* Once a synonym for "artistic", but that's a distracting coincidence

...

Well you're not replying, so I'll respond to my own "story" about what you meant, as if I guessed right. I think art can be lots of things, and this (interpretation of cryptic output) is indeed one of the things it can be. The aesthetic movement in the 19th century opposed didactic art, such as deliberate allegories, hence they coined "art for art's sake". But this shows that art as deliberate communication evidently existed too, since they rejected it. Then, looking somewhere much more modern, pop music artists frequently surf atop of fictional images of the star that are largely invented by the audience. Mystique is quite enjoyable. But wanting to be fooled, wanting to misunderstand, is perverse. Milli Vanilli's fraud pissed people off, and rightly so.

There was probably some pomo shit about completely interpretable art, but I doubt anybody really liked that.

So I think the popular preference is for art that is inexplicit, somewhat mysterious, certainly not a plain communication, but art that is an effective inexplicit communication and has a genuine artist's mind invested in it.

I could add that this is somewhat preserved in art even after it's been digested by AI, but that's hardly ideal.

The point is that human-created work has more meaning, because it's the product of a real life. Writing is a reflection of the writer.

Computer code can replicate rhetoric and prose. It can even create meaning the user didn't anticipate. This is as much a reflection of the 'author' as any hallucination.

The kind of essentialism people hold between the object and the meaning of the object is effectively inseparable, and it's pointless to argue that it's not or shouldn't be. I write in the golf space. There have been many, many replica golf courses built, and they have been largely unsuccessful, except one that is barely open to the public. The have been mostly unsuccessful not because they don't replicate the originals well, but that the vast majority of folks don't want a facsimile, even if it's fine.

I don't live like that, you don't live like that, but most people do, and they have ever right to. Humans aren't analytical engines. Evolutionary biology has pushed us strongly towards caring deeply about meaning, identity, and respect. This is normal and has to be unlearned. Good luck with that. Other people don't have to like and respect art for you to.

Except AI art is not a exact copy of existing art, it may sometimes occur, but most AI art is people who are not artists expressing themselves. Its a specific personal vision of a golf course only for themselves in most cases. Its irrelevant who else likes it. Previously even the lowest tier of contract artists on websites like DeviantArt would ask 20-40 dollars for such a thing, and that's after you first searched around for an artist who does your specific idea/genre/subject. Then there's the back and forth, waiting in a queue, and potentially taking money and ghosting and so on. And finally how few people have 30 dollars spare cash to blow on a throwaway idea?

Why is it "bad for art" if someone dislikes AI art then?

I don't know, ask the people who think AI art is bad art or not art.

That wasn't my question. The poster above said that snobbery towards AI art is "bad for art" and since you supported their general pro-AI arguments, I assumed you were the same poster. Sorry, I didn't look closer.

So do you think AI is OK for throwaway stuff but not OK for "production" so to speak?

Its okay for anything. I don't care if the pixels and vectors for a character or landscape in Skyrim was made by humans completely, by humans in heavy collaboration with AI or almost completely autonomous AI. For some art eg live music I would prefer a disclosure, but that's not because I hate the use of AI rather it is false marketing.

So why is it "irrelevant who else likes it"? If it's made for the public and the public doesn't like it, then it's pretty relevant. What if people do care how the Skyrim character models were made? What if they do hate once they learn it's AI?

People can have any reasons they want no matter how insane. By revealed preferences, the split has already occurred between those who hate AI and those who don't mind AI art.

That's not the answer to my question but fine, why is the split important in any way? Some people hate jazz, some people love jazz, and it follows that... what?

People are generally emotionally weak and hate to think in ways they are in fact simple physical systems and not special. Humans have been knocked down a peg several times over the history by natural sciences. Nothing new.

There is also the generational factor. Digital art, printing press etc all of it had a generational factor to it before gaining acceptance. So who knows what the future holds.

I see. What do you think people seek in art? Why do they engage with it?

For most art they just like the picture, caring about the artist is inasmuch as one was required to make it. Already before AI, if I selected a random photo from a given genre in DeviantArt or ArtStation it would be hard to make deep philosophical queries about the artist like the original starter of this thread tree was thinking is necessary for art. Most of it was surfaced when people searched for xyz wallpaper, they click Save and set it on their phone or desktop.

Even for pure picture lovers, I will say beyond philosophizing about the creator, the content of the picture itself is subject of thought and analysis as far or as little as one wishes.

People care a bit more about the creator in music, because musicians have built up a lot of image and fake mystique. But then again, for some genres, its much less relevant to people especially those that got grew natively or thrived on the internet eg noise rock, hardcore electronic.

Some of course want deeper analysis, seeing the artists blog/thoughts etc.

Ultimately the actual revealed preference is that most people just want "pretty pictures" and couldn't care much about anything else.

Interesting. I am one of the people who adores visuals. Instagram and Tumblr used to be my favorite places, I had my own pretty pictures curation blogs, my way of relaxing was creating Pinterest boards. Now, I also like "serious" art but I wouldn't scroll museum galleries when I just needed to feel a little whimsy, I went to Pinterest and often didn't care about the author behind them in many cases because the pictures were derivative enough to not feel "special" (but sometimes I would and then I would discover new artists).

I don't engage with this hobby anymore since the platforms got flooded with AI-generated content. Now, I do like some early and cheeky AI-generated images (mostly pre-ChatGPT), but not after -- and not because of corporate and social politics -- I'm honestly largely out of the news cycle and don't know what terrible or amazing thing some AI company did this week. And yet my hobby completely died and it was strangely disorienting because I struggled to find something else that could replace it, I don't have social media and didn't have feeds except for on those three platforms (I deleted them all now).

I am an outlier from the general population in how I engage with visuals, and I do think that it's likely that all of the commercial visuals in the future will be AI generated, now, I don't think they will be pretty, though. I think they will be largely ugly. I think the majority of people don't actually care about "pretty pictures" at all and are content with very ugly things surrounding them all the time. And the rest can't help it.

I don't mind ai at all, its the same for me except occasionally I get to express my own idea. Prompting is still not easy for getting a very specific idea so wanting art as a proxy for someone having done the work, I don't mind finding someone's AI art for a specific composition or idea. For stuff that was made with generic prompts too, as I said before, it is like seeing the AI's vision and thought process. I don't mind either way. Again even if a full AAA game was made with AI as long as the result looks coherent and nice I don't mind.

Case in point, I guess.

I don't find it ugly. Its equal to average or better than the low end of wallpapercore. If you think AI is ugly in this specific genre then its your bias clouding your thinking. If AI art in this genre is bad then most of DeviantArt and ArtStation was also bad.

I wish time machines existed so we could do blind testing of AI generated art vs purely manual art to people around say 2019.

I don't doubt you don't. Why is my judgement clouded and not yours? DA was p bad, it was for amateurs so it was expected. Never had a chance to use AS.

If you want to have a discussion about "what art is" I think I would end up arguing that people creating AI art are artists.

My point is it doesn't matter. All art ultimately boils down to is doing something cool, and then pointing at it. The vast, vast, vast majority of "art" is not paid and never was. Even the art that is called "art" has little-to-nothing to do with effort. Duchamp's "readymades" are basically less effort that typing a prompt into AI.

That said, nobody has to like Duchamp's readymades. Nobody has to like anyone's art. The point of art is that you can like it or not. It's not objective, but the vast majority of people are going to pretend that it is, but that's fine because it's actually not.

Another point is tat I think current LLMs are arguably conscious. I am not saying they are but I am open to the possibility as human brains are physical systems that are conscious so there is no reason other conscious physical systems might not exist. In that light, or even otherwise, I think if we avoid heavy handed prompting or micromanaging, the AI itself has its own ways of thought, tics and style. So in that sense we can cite the AI itself as part composer.

There are plenty of human consciousnesses whose output people can't stand. If AI is conscious, it's perfectly in line with how human art works -- sometimes someone's creative output is offputting to a great deal of people.

Under that premise we would consider a photography by my 6-year old a greater achievement than a detailed realistic painting, which we still marvel at.

I think that, as with photography and painting, there will be a place for all types of art, also machine generated, but under different umbrellas.

To think that they compete in the same category is a mistake.

I discuss it a lot with my wife who happens to be a painter.

For years we have perceived my work (swe) as a safe harbor. With AI boom, tides have changed and now I perceive her work as a much, much more future-proof activity

I agree with you, people want art made by humans, I don't think most people care if the code was written by humans in the same way.

While I agree that people _want_ art by humans, they mostly buy the cheapest possible thing that fits their taste and intention. Take furniture as an example of art component pushed to the premium segment

I think furniture is probably closer to software than art for most people but I know what you mean.

There is evidence that intrinsically motivated communities of programmers care a lot about human code.

Some open source projects have already banned AI contributions. I think this trend will continue (in stark contrast to the corporate world).

I must admit the prevalence of AI generated images has challenged me to be more attentive to detail, I even started looking for small artists I can commission for stuff on my walls, been going to expos a little more... I'm happy if this gives artists a small lifeline away from the more mercantile aspects that AI is conquering.

That's been the case for over a century. People pay a large premium for the original even when identical copies are available.

A lot of us saw this one coming a while ago. Many, many artists of all kinds have already put up "AI-free" (or similar) badges on their work. I think it will be more significant in the fine arts world, however local art fairs could be affected as well.

I say the more the merrier, but artists who use AI in their workflows are going to be pressed about it - with negative connotations - for a while. Transparency is the best solution in my opinion.

People trying to use AI in their creative process but not as their creative process might be the biggest losers for a while. Right now there are a lot of negative gut reactions to having any AI associated with art at all

Losing implies that there was something to lose to begin with. AI as a tool for art is new and there hasn’t been an artist who has managed to do something with it to break the internia.

Art is not ‘supposed’ to be easy to do, if one believes that they see something important they can express via AI — they are free to do so and try to say smth to the world, AI tools are relatively accessible. It’s not guaranteed that the world will listen now, but that has always been the case — many great artists never got recognition during their lifetimes including those who worked in novel paradigms.

So no, they’re not losing anything. If the absence of the immediate mass approval of their art is the obstacle, well, then they don’t need to make that art.

Why does how hard something was to make have anything to do with how good the art is? Three chord punk bands were in my view artistically superior to a lot of rock music from before as they made a structural innovation with the power chord.

It doesn’t. Art isn’t ‘supposed’ to be easy or hard, it is also not ‘supposed’ to be economically viable or non-viable. Creating art was all of these things (hard, easy, profitable etc) in different places and different times. Art is still very hard because art is an expression of a human finding something true about the world they want to pass on. The tools don’t matter. Public recognition, largely, also doesn’t matter for good art (as long as it speaks to someone.) AI tools are easy but public recognition is hard to obtain. Both of these things have nothing to do with the quality of art.

If an artist has something to say with AI, they can just say it, no one is going to imprison them. Artists did that throughout the entire history (sometime even with the threat of being jailed), nothing changed in this regard with AI.

Art is hard not because of the tools but because you have to have something worthwhile to say (most people don’t) and you have to figure out a way how to get another human to listen, and also you have to hope that someone some day will listen because what you’re saying is true, even if it doesn’t sell today or ever.

Commercial and pop art has always existed also. I think it is good as it brings joy to millions of people who don't have the time or money to commission artists.

Sure. If the gripe is that it’s difficult to make money off of AI art, well, it’s difficult to make money off of any art, especially if someone is working with unusual materials and tools. The market will decide if there is demand for it, but the market in general has historically been an extremely poor mechanism for matching artists with the audience who can see and appreciate their art. Art makes a poor product. The starving artist stereotype exists for a reason.

P.s. Pop art actually directly talked about this relationship with the market and it’s interesting because of that, not only because the colors and shapes were fun to look at (but that too).

Why aren't we seeing a premium soft "Software made by Humans"?

The software part is quickly being solved but I honestly believe the more creative aspects will still see large demand. LLMs have a certain stink to their UI, being creative is a hard skill, which AI has yet to master.

For so-called "AI" to have creativity, we'd first have to work out what that even means for ourselves right? What is creativity in the first place? I'm unconvinced it's just turning up the "temperature" or fiddling with some pertubing parameters.

EDIT: And AI is also pretty terrible at being consistent too, you have to work quite hard to get a level of "consistency" that you're happy with.

Most of what gets called creative isn't creative at all. Often it means "Produce a content product which approximately matches commercial genre expectations with just enough of a twist to stand out a little."

A lot of people can do that. A lot more people could do it if they tried to.

People who are genuinely out of the box creatives, with a unique blend of original vision, technical skill, emotional and psychological insight, and genre-changing taste, are not common.

> "The software part is quickly being solved"

There will always be problems to solve with building software, regardless of the way it is written.

From code to no-code to vibe-code, to even more code. What we are seeing here is more of the same being built.

An example of this is "sandboxes". Everyone was writing "sandboxes" until it there was 1 vulnerability needed to break them and we will keep seeing this with developers vibe-coding infrastructure.

A reply quite some time later, but you are definitely on point, I rescind my previous statement. At least, I do think you are correct about always needing software or having problems to solve in it. Still undecided on the creativity stuff.

Because that's not the point. Before you could waste, without kidding, hours in meetings discussing css alignments, flows that didn't make sense but the upper levels were pretty serious about it - "Our QA and test groups say this is of the utmost importance", even when the QA and test groups were the same, 3 or 4 people that insisted on it. I think most people, but at least I, could always deliver things inside schedules (not even set by myself) without a problem, even with those "adjustments" and changes of mind, and doubly so if I could just have the same leeway that now seems to be in place.

Nonetheless, if there's going to be a premium it has to be discernible in any particular way because software for business is always considered a cost. But what really shows this industry is half a joke, is that now there's no issue in giving 3rd party agents full access to inner code and logic, many times credentials too because these systems are being deployed in CIs. And because there's really no enforcement or punishment for leaks, or endangering users in most areas there's no reason for a premium. There's people pushing thousands of lines / min (and I think even the most professional people out there can't honestly keep up with it) although that isn't in itself meaning the overall security is worse than before, it just means that for the most part it's now un-auditable in any meaningful way.

Because code is not art and the corporate culture doesn't care about artisan code, neither do the users. As long as the software sells, they are happy.

In fact executives demand the opposite: token max to get the features out of the door faster, all in the same of closing that one last sale.

I do prefer OSS over proprietary software because to me, at least prior to GenAI, OSS was the label that was "software made by Humans" where proprietary was "made by corporations". It didn't make OSS win commercially, of course, and I personally don't think that human-made art will win commercially.

I’ve started to see (and react negatively) to a kind of “Claude design”. Apps, slides, images share the same aesthetic, and even after aggressive prompting it reverts to a the mean.

I even create these things. But that aesthetic has started to imply a kind of throwaway, low value thing

What will be interesting is if models can create more diverse styles than the neutral Claude default we see everywhere

I think that this is also a matter of organisation that people can get better at if they understand a bit of how things come into place. If you start by creating a stable theme, with the spacings you want, the colors you want, fonts, sizings, etc, even using an AI, then you can "customise" in full the look of a web-app.

The html there's not really many places to be creative, a navbar with 3 sections, 5 links, a text card, a hero image, etc, all those are the same across thousands of projects - they're structured, there's not really many ways you can be creative about it. So in part seeing that "middle ground" design is indeed a filter for the lowest of the low effort (or not understanding how to prompt properly, or how the final thing comes together).

If you just want to prompt models and have them come up with more diverse styles that's also quite, or more, lazy - it would help if you understood (not you in particular) how the look of a web page comes together, because then you can direct the model properly - and if needed tweak the lose ends because at some point it's just easier and faster to edit the output yourself but you need to know what needs change and how. If you know how an html page comes together into the thing you see in the browser you can spec the model properly.

As a human artist/film director/animator/musician etc I view "AI" as just another tool in my toolbox and what a tool it is!

In animation/film production- studios would have a "TD" on hand- a technical director- a programmer on staff to tweak workflows for artists, build new tools etc-

Now my TD chills in VsCode waiting and working for me 24/7 in the form of Claude Code, Kimi Code etc

I can also have my TD embed another TD(Deepseek flash 4.1) into the new tools(making them smarter and more useful) my TD builds for me while I am in active production on the same PC

All of this allows savvy creatives to dream bigger as a single person is capable of anything a "big studio" can do-

So artists need to leverage these new tools and be more ambitious.

Aesthetics are dead- original ideas are it.

There's one thing worse than "designed by AI" and it's "designed by committee". I wish more buildings were designed by AI. There's so much depressing architecture.

If the AI is tasked with "Build a building as cheap as possible with as much floor area as is legally permissible", then it too will make dull buildings.

What you really need is more people willing to pay a premium to live/work in a unique or famous building like the empire state building, and then more such buildings will be built.

That is a good point. However, from time to time we do spend very large sums on certain buildings and what we get for it isn't proportionally more beautiful; in fact it often looks worse.

Hopefully with AI we will make things like stonework and indoor plasterwork, hell, even indoor statues, that much cheaper, and we can start creating beautiful places to live.

one person + ai >> a committee of people

And consequently a gray area of generated slop trying to pass off as created by people.

Feels like approximately 80-90% of AI generated art, writing, games, tools, etc come under that category now. Very few people seem willing to come out and say "yeah, I got Claude/ChatGPT/Codex/Gemini to make this".

Or if they do, they'll try and downplay by saying they used AI for some elements, but not most of them.

Been a big problem in music.

On my site, we started asking artists to disclose their use of AI starting ~2 years ago. Turns out roughly 1 in 3 lied.

So, I developed an AI detector. It's working pretty well for that. But it's a schlep having to keep up with all the new models. We're processing ~60,000 songs a day at this point, kinda crazy!

How do you handle false positives in the detector?

We have an appeal process. FP rate is super low - accuracy around 99.4%.

> Been a big problem in music. On my site, we started asking artists

What’s your site?

SubmitHub

Curious on two things

1. How often do you encounter false positives, and does that change things?

2. Have you considered outright prohibiting/banning vs disclosure?

Just want to hear from someone running a platform how they deal with this since I imagine it's very hard to get right.

1. Rare. 99.4% accurate. 2. Yes, in September we outright banned slop. Got a lot of backlash for that, but I think it was for the better.

I can imagine that being a reaction to the negativity and backlash from a substantial (or very active) part of the audience, to anything built with AI. I don't imagine that shame for using AI for something, or fear of disclosing it, is self-imposed.

Ideally, how you create something and what tools you use should be nobody's business, and all it matters would be if they like it or not. But inherent value / merit is a thing, and concern for the provenance of what you buy is also a valid position for the public to take. It's gonna be a while before the balance settles on either AI = bad or AI = who cares.

As AI art spreads, provenence and human intent may become the real markers of value.

cheating in chess is illustrative. the output is almost identical so we require proof of human effort. nobody wants to watch computers play.

Similar trajectory to "human premium" for human-authored text (when we started using synthetic AI-generated text in pre-training, around 2023 for maybe llama-3?). AI text is OK (kind of like a low nutrient content food) but labs still pay a premium for high quality domain-specific data silos. As far as the consumer, pattern-breaking media will always have a place, slop videos follow a similar distribution and so there's naturally pushback against seeing the same thing, over and over even if it's not inherently crap (just as in AI-authored text).

While there is a lot of slop, it is possible to create art with AI that does not look like slop and is as creative as a human creation. That's really the main cool thing about AI is you can create things you otherwise couldn't, and the better you are the higher the ceiling for the ai art is. It is a bit of a struggle though, as in prompting and finding the right references is difficult to get good results

Edit: here are some examples for those wondering what I mean: https://imgur.com/a/mK3Iuzq

You can sort of tell they are AI because of the composition stuff but I think they are aight.

I don't see how AI was necessary to create these.

It's not needed for a professional designer / photographer etc but now a person who does not have those resources can create them

Why would you want to? These are advertising for a luxury goods company. You have to include the actual goods in the picture, not a hallucination, and you can afford to do it properly.

And any computer program can be written in assembly, so why have other programming languages?

This is comparing apples to oranges.

My point is that "X isn't necessary to create Y" is not a useful critique of X or Y, or of the process of creating Y using X

The killian and polo images are strange. There is a mirror over an sofa in a poor lit room. I said strange, not impossible.

I agree with your thesis, but your link is to marketing images not art.

That's fair but there is a lot of creativity in prompting and finding the right image because on the way to those images there is a lot of experimentation and I came up with the ideas for the images not the AI.

Creativity != art.

The issue isn’t the quality or the visual result. It’s not just slop because it looks bad.

Art is as much about the journey as it is about the destination. It’s the friction of someone going through each detail, making a choice. It’s the artist taking time to figure out how they want to convey what they’re feeling and thinking.

The reduction of this friction culminates in the reduction of the piece’s essence.

Yeah I understand your point because I've made music and design for 25 years the old "manual way", and there is definitely less friction with AI than there is the manual way.

But what interests me about AI is that it is possible to do things that you couldn't do before particularly in processing images. For these images I took photos of the bottles myself and used other images and the AI can place the camera anywhere in a scene and place objects anywhere in the scene, and things like that. You can create almost any image you can imagine with the right sources and the right prompt, and even here there is some friction because many images do end up looking like slop and you have to prompt it right to get something non-slop like out of it (though not always).

This ability to manipulate and process information is what makes AI interesting to me, although I'm not sure I adequately explained it. It feels like a new tool like Photoshop except across a different dimension. I feel like we've barely scratched the surface of what's possible to create.

I’m going to be honest. Those look like shit.

Another area where prices have increased due to AI because a bunch of soulless chronically online people can feel better about themselves.

Mind you, human-painted artwork is a niche beyond expensive art. Not everyone buys the equivalent of Monet paintings. I've been eyeing a painting by a local painter just to realize the prices jumped from $500 to something like $3000 on average. Sucks.

How are the prices a problem? If a person is just looking for a picture not a particular artist or philosophical theories, AI is cheaper. If a person wants to commission an artist, that was always a semi luxury even for cheapest of cheap rate online artists.

In fact if your old price was already 500 dollars then its already a useless luxury. People think a lot before paying 500 even for actual tools, anyone who has 500 to spend on a non-essential luxury is already doing something most people don't have any need to.

Are you saying that artists getting payed more is a bad thing?

I'm not sure what you're saying. Is the claim that the artist raised the cost of his work 6x because he's getting less work, because people are using AI instead? Maybe I'm in a bubble, but that doesn't seem credible. I'm trying to imagine the person that was thinking that they would hire a local artist to paint them a painting and then discovered AI and instead asked Gemini to do it. Those just seem like two completely separate use cases.

After reading multuple discussions in this thread and a bunch of others, I feel like people on HN think about art skills in the same way as they think of their software skills.

Correct me if I'm wrong, but skills that allowed one to make software largely made one pretty employable in the last few decades.

Now, art skills do not make you employable. There are narrow industries where you can make money but throughout the entirety of human history, being an artist rarely provided any sort of economic security. Some artists did and do make money. Sometimes it was prestigious but the barrier of entry was insanely high. And yes, getting to their level of skill and sometimes even getting access to the tools (look up how much a decent harp costs) was yet another enormous obstacle. Out of the entire pool of people who had artistic abilities, very few got access to tools and resources to be able to express their abilities. Out of those, even fewer could make money off of it while also doing artistic work that was deeply fulfilling and aligned with their talent. Occasionally, the stars would align and you would be lucky to be born in a time and place where you could get the schooling, the tools, the job that paid well enough and that didn't makwe your soul die because it alined with your natural creativity.

There are a lot of people who (I think) speak of AI as if democratizes access to the tools and therefore in their mind it should follow that there must be some market for their newlyfound creative expression. But if you ever seriously considered art as a career, you know that you will likely not have a market for your skills. It is difficult to make money even as a traditonal artist (as in a painter) or, say, a musician, god help you if you want to do mixed media, if you want to play avant garde jazz, prog rock, or Barocco music, if you want to be an actor, a playwright or a ballet dancer.

Some people (I assume those who are excited about the idea of being able to express themselves) seem to be bummed that the world rejects AI art as 'slop'. I am not going to say anything about effort, laziness, etc, I believe anything can be used to make art, even a toilet put in an exhibition hall (I like Duchamp, I like modern art, I like post-modern art, I like whatever-the-paradigm-we're-in-now-art -- metamodernism? I love art, I did see early pre-ChatGPT experiments with neural networks that moved me), but I have to say something that any aspiring artist must know -- you will not make money off your art. The people who will be making the 'human premium' that the article talks about, will be few, just like there are few of them now. AI artists will not make money of their art. Art is a terrible career choice, now and tomorrow.

It's still worth pursuing, in my opinion (as I said -- I love art), but there is little point of talking about its commercial viability. It will always be shit.

I disagree. I think the result, the quality, is what matters, not how it was made.

Around 2010 I made music videos for a Vocaloid producer. Back then, people sometimes said Hatsune Miku songs were worth less because a machine was singing. I didn't think so, and in the end the songs were accepted because they were good.

Of course, Vocaloid songs were still written by people, which is different from AI making the whole song. Even so, I think what matters to listeners is the quality.

(I wrote this in Japanese and used AI to translate it.)

We should just admit that human created art is dead. Just like hand coded code. Same thing. AI will soon beat any human in quantity and quality. Just accept it. It is over for humans.

Just like photography killed realistic paintings and vynil killed the need of live music.

And soon, companionship, emotion, community, and sex!

We didn't need AI to destroy those

Code has always been seen as purely utilitarian by most people, art is different.

I suspect use of AI to "realize one's artistic vision" will get increasingly more normalized, but I don't know how people will react to end-to-end AI art (which generally doesn't exist yet), things are going to be a really strange once/if human taste gets fully mapped and demystified.

Reply on news.ycombinator.com