Procreate defies AI trend, pledges “no generative AI” in its illustration app

Procreate CEO: "I really f—ing hate generative AI."

See full article...

 

[iindigo]

But look at Adobe. It is more popular than ever. More polarizing too because of deep AI integration.

People have been looking for ways to escape Adobe for ages now, even excluding any AI shenanigans. It's actually been an ongoing arc for over a decade, starting when Adobe shuttered Fireworks after buying Macromedia without equipping Photoshop or Illustrator to be proper replacements. That's what gave rise to both Sketch and Figma, with the former of those two in particular taking tons of cues from Fireworks.

Going subscription-only was the second big push, which has created a market for the Affinity suite and Pixelmator and bolstered sales for Clip Studio Paint. It's also been a factor in increased donations to the Krita project, which has helped it become a pretty respectable FOSS graphics editor (it's not just GIMP anymore).

Most recently the "clarification" had to make that they don't use user data to train with was another goofup. Regardless of their ambitions in the ML space surely someone there had to have a clue that not making the EULA clear about that was stupid.

 

[hizonner]

"I want this person's arm here, not there, and the background to be composed like this, and the shadows to do that" - it is possible already, for quite some time. It's called ControlNet. And the new models are very good with prompt adherence so stuff gets easier all the time.

That's steering with images... and it has a significant learning curve. And it starts to look like a fairly unpleasant form of work.

"from where she happens to be sitting right this second, she could stick out her arm and pick up a pencil and paper" - you do understand that this personal anecdote doesn't change accessibility of art for an average person? She is not typical.

Neither is literally anybody who has ever used a generative image model. People who've done that are absolutely a minority among humans.

In fact, I suspect that if you could count the number of people in the world, right this minute, who can do a recognizable hand drawing of a specific person's face, it would significantly exceed the number of people in the world who have ever interacted with such a model even at the crudest level. Even though I personally am in the second group, but not in the first.

You also made this about it being engaging, and I very much doubt you have anything at all to support the idea that the "typical" person finds the AI more engaging. I wouldn't want to make any guesses myself.

Anyway, the point I made, the point that you've supposedly been arguing against, was that what's accessible depends on who you are and what you already know... which means that what's "typical" isn't really central to the whole thing. You seem to want to deny that anybody could really prefer working manually, but when presented with a counterexample, you fall back to an (unsupported) claim about what's "typical".

And even for her - the skill she has has no market value.

No. Skills. Will. Have. Market. Value. In. Fifty. Years.

None.

If you're much younger than I am, and you stick to this obsession with market value, then you're likely to have a very unhappy time in a few decades. The market is not going to care at all about what you or any human knows or can do, because AI will be able to do all of it better, with no steering from anybody, and at a price below what it costs to buy the food to keep a human alive. So hope that the market value of your skills isn't what determines whether you get fed.

 

[Aurich]

The image I linked before was specifically in response to your claim that "it's just a prompt". There was a workflow description. The picture took the author 6 to 8 hours to complete in multiple programs. And it was actually awesome:

Yesss, girl on a dragon.

 

[hizonner]

Yesss, girl on a dragon.

Dude, your avatar is a skull. If we want to talk about banal and cliched, I mean.

The subject matter of Tall Dwarf's picture isn't electrifyingly original, but from a technical point of view it's leaps and bounds beyond the straw man you claimed was equivalent.

 

[MidnightHacker]

I only dabble and may not even use the software much, but you sir today made a sale. I appreciate you for saying what needed to be said!

 

[vpointcon0011]

When I could have some GenAI "draw" it in a second?!? /S

Spoiler: one of the less unhinged genAI results, but still giving me some nightmares

View attachment 88355

There is something unsettling about those flowers.

 

[Marcus Andreus]

I know this is a tech site—and we often take fairly exotic tech ideas for granted—but it is utterly unhinged that anyone here is trying to argue that learning to prompt an AI is in any way a more accessible method of creating images than a pencil an paper.

Utterly unhinged. Completely detached from any semblance of reality. Like trying to assert that yacht racing is a more accessible sport than soccer.

Once you have a pencil and paper, that's all you need. You can draw. You have accessed drawing. Everything else is optional. The minimum tool for creating AI images is "computing device." That is literally orders of magnitude less accessible.

I went to the KAWS exhibit at the Art Gallery of Ontario a few weeks ago, and one of the most unexpectedly fascinating things to me were some of the pen sketches and the way he creates the idea of heavier lines in them. Sure, he has lots of other media available to him, but he can still get stuff done in interesting ways with materials that people will literally give away.

 

[OrvGull]

But look at Adobe. It is more popular than ever.

That's because industry lock-in forces people to use it. Adobe could put "artists suck" on the splash screen and people would still buy it, because they have to.

 

[randomcat]

That's because industry lock-in forces people to use it. Adobe could put "artists suck" on the splash screen and people would still buy it, because they have to.

And yet rivals like Procreate struggle for a fraction of the market share on one single device. They hear "we wish someone was as good as Adobe, we'd give them money in a heartbeat" and fail to understand that this means they are empirically not as good as Adobe.

Another fired-up CEO video where he uses the F-Word ought to do the trick! We only have one example in this thread of a Procreate document and it looks like MS Paint from thirty years ago, so feature demos sure ain't gonna do it...

 

[OrvGull]

And yet rivals like Procreate struggle for a fraction of the market share on one single device. They hear "we wish someone was as good as Adobe, we'd give them money in a heartbeat" and fail to understand that this means they are empirically not as good as Adobe.

Procreate is actually very good, but it targets a narrower set of tasks than Adobe Photoshop. It's not intended to be a photo editor. It's intended to mimic drawing on physical media. It excels at that, and for people whose work involves drawing original art it can be a professional-level tool, but it doesn't have the broad toolbox of abilities you get with Adobe Photoshop. On the other hand, you could buy it outright four times over for what it costs to rent Photoshop for one month.

But for commercial illustration work it's impossible for anyone to displace Photoshop, because only Photoshop can guarantee the ability to read any PSD file a client sends you.

 

[randomcat]

But for commercial illustration work it's impossible for anyone to displace Photoshop, because only Photoshop can guarantee the ability to read any PSD file a client sends you.

Careful now, you're close to bursting the bubble. There are a lot of people here desperate to believe that if not for this pesky AI business they could somehow make a fortune doodling on DeviantArt. They don't want to hear anything about what making a living as an artist is actually like, that would ruin the fantasy.

 

[JoHBE]

Sometimes you don't want art. You just want a picture of X. Without investing any more time or money than necessary.

It's a trap, not dissimilar to how we keep using fossil fuels in many cases because they are "cheaper". They aren't, in the long run, because the real cost is invisible/externalized. Just like generative AI has "externalized" its apparent cheapness, which eventually relies on stealing. But will end up dissolving the very fundaments it is standing on (human artistic output and innovation).

 

[Tall Dwarf]

You really want this to be able some straw man. What adobe does, or which skills are marketable has absolutely nothing to do with what I am saying.

Please actually read what I said, this is getting silly. You even went back and quoted it, and still failed to understand it.

No, I didn't fail. Your little scenario is just a bunch of delusional lies, just as I said: "that's just not what happened." No strawmaning here, I didn't put words in your mouth.

Lets break down what I said, sentance by sentance
This is an example of a product designer not doing the research necessary to understand what their market really wants. Instead they make a superficial assumption (that joggers have a transportation problem), and failing to understand what their real motivation is (they enjoy jogging).

No, this is not. Tens of millions of people are using generative AI. Lack of interest from the market is not a problem at all. Adobe is doing just fine and new tools are clearly popular enough to keep investing into it.

People who enjoy jogging can keep jogging. Believe it or not, you are not the center of the Universe and there are other people in it, with vastly different demands and interests.

At this point they are going ahead and building a product based on their bad assumptions. Making an investment in building something their target market is not interested in buying.

What are you talking about? AI image generation is wildly popular. While there is a noisy opposition, it is already normalized. It's in Photoshop, in MS Paint, in X. There are solutions of all levels, from people making stuff in browsers to enthusiasts building PC specifically to work seriously with AI.

The joggers yell at the sky "they took our jobs" because those who don't want to adapt left behind.

Now we are at the end of the product cycle. Jogger are not the ones raging here, they just have no interest in a product that is useless to them. The person raging here is the bad product designer who made a product for the wrong market. They are the ones struggling to understand why no joggers are buying it.

"end of the product cycle"? This thing just started and it develops rapidly. This month FLUX.1 was released, a high quality model with public weights. Community is wildly enthusiastic about it. About a week ago it was added to X.

No, no one is "struggling to understand why no joggers are buying it" because no one cares. Plenty of people buy the new products.

And you are lying - the joggers are "the ones raging here". They are being replaced and they don't like it even a little bit.

Your entire "joggers" story is a bunch of delusional lies. It is not rooted in reality in any way.

And no, by definition joggers cannot ride, sit or go back. If they don't jog they are not joggers. Eg people who ride are called riders, not joggers.

No one cares if they keep jogging, rage quit or get kicked from their jobs by competitors. A lot of them will stop jogging very soon.

 

[Tall Dwarf]

It's a trap, not dissimilar to how we keep using fossil fuels in many cases because they are "cheaper". They aren't, in the long run, because the real cost is invisible/externalized. Just like generative AI has "externalized" its apparent cheapness, which eventually relies on stealing. But will end up dissolving the very fundaments it is standing on (human artistic output and innovation).

"Externalised" as in "it allows new people to compete in the market". Which is not quite as tragic as global warming.

"But will end up dissolving the very fundaments it is standing on" - or not. The whole thing is 2 years old. It is quite possible that training doesn't need billions of images to succeed, this is just how the current generation of tools work.

And it always will be possible to pick what AI is trained on. The success of Midjourney is exactly in using people's artistic preferences to tune their AI to fit them better.

FLUX.1 has 3 models - "pro" is the real deal trained on images. "dev" and "schnell" are "distilled" models - they are intentionally trained both on original data and the output of "pro" model. And they are still awesome.

Your pet theory that AI will eventually collapse on itself is just that, a theory. AI generation has the only way, up. Because if a new model is inferior it is scrapped.

 

[randomcat]

"Art is a struggle, it's a journey, you need to devote time and effort to it, you need to put in the work and learn the craft"

types "boobz" in the prompt of the cheapest, easiest SD frontend and clicks Generate with no additional settings, is dissatisfied with the result, immediately gives up

"this GenAI stuff is bullshit"

 

[Tall Dwarf]

That's steering with images... and it has a significant learning curve. And it starts to look like a fairly unpleasant form of work.


Neither is literally anybody who has ever used a generative image model. People who've done that are absolutely a minority among humans.

In fact, I suspect that if you could count the number of people in the world, right this minute, who can do a recognizable hand drawing of a specific person's face, it would significantly exceed the number of people in the world who have ever interacted with such a model even at the crudest level. Even though I personally am in the second group, but not in the first.

You also made this about it being engaging, and I very much doubt you have anything at all to support the idea that the "typical" person finds the AI more engaging. I wouldn't want to make any guesses myself.

Anyway, the point I made, the point that you've supposedly been arguing against, was that what's accessible depends on who you are and what you already know... which means that what's "typical" isn't really central to the whole thing. You seem to want to deny that anybody could really prefer working manually, but when presented with a counterexample, you fall back to an (unsupported) claim about what's "typical".


No. Skills. Will. Have. Market. Value. In. Fifty. Years.

None.

If you're much younger than I am, and you stick to this obsession with market value, then you're likely to have a very unhappy time in a few decades. The market is not going to care at all about what you or any human knows or can do, because AI will be able to do all of it better, with no steering from anybody, and at a price below what it costs to buy the food to keep a human alive. So hope that the market value of your skills isn't what determines whether you get fed.

"That's steering with images... and it has a significant learning curve" - non trivial things are non trivial, who'd thunk that!

"And it starts to look like a fairly unpleasant form of work" - an opinion. People are doing this stuff in their free time because it's pleasant to them. Your tastes are not universal. Big surprise, I know.

Neither is literally anybody who has ever used a generative image model. People who've done that are absolutely a minority among humans.

We are talking about accessibility in US. You are trying to extend it because you are intellectually dishonest.

For a person who wants to start drawing AI art is much more accessible because it doesn't require years of practice. There are plenty of free sites that allow to generate images and this number will grow as technology gets better. Just last week it wadded to X. Basically, all you need is a smartphone to start using it.

"In fact, I suspect that if you could count the number" - see, if you start with "in fact" there should be some actual facts, not your suspicion. Because that's not here nor there.

You also made this about it being engaging, and I very much doubt you have anything at all to support the idea that the "typical" person finds the AI more engaging. I wouldn't want to make any guesses myself.

The fact that people enjoy creating AI art is obvious,. The fact that it's trivial to start using and achieve good results is obvious. The fact that millions of people who didn't make pictures before make them now is obvious.

Yes, it is obviously more engaging. Otherwise all these people would be sitting with their pens and papers and no one would ever notice them.

Anyway, the point I made, the point that you've supposedly been arguing against, was that what's accessible depends on who you are and what you already know... which means that what's "typical" isn't really central to the whole thing.

That's not a point, that's just ridiculous. In order to talk about accessibility you have to use generalizations. That's the only way. "There is no typical" is a ridiculous cop out. Your personal anecdote about your daughter is completely irrelevant. And it is not even clear how much truth in it.

Tens of millions of Americans used generative AI. And this number rapidly grows. It is impossible for every single one of tens of millions of people to be non-typical.

No. Skills. Will. Have. Market. Value. In. Fifty. Years.

You. Don't. Know. That. So. Stop. This. Pretentious. Crap.

And your daughter already determined that her pen and paper skills are a dead end.

If you're much younger than I am, and you stick to this obsession with market value, then you're likely to have a very unhappy time in a few decades.

Whoa. You made wild assumptions and predicted my future based on it. Did you use crystal ball?

So hope that the market value of your skills isn't what determines whether you get fed.

Don't worry about me, I am fine. Also, did I ever ask for your wisdom? I am fairly certain I did not vbecause I find your ramblings rather funny.

 

[snowram]

ProCreate seems to support addons. If anyone want generative AI in there, it should be relatively trivial to make one by hooking up an instance of Automatic1111 or ComfyUI. I have already used that kind of addons in similar softwares with great success.

 

[Tall Dwarf]

Yesss, girl on a dragon.

You sneering because you can't really argue with facts.

Your "you can pretend it was any one of them" and "Here, close enough" are obvious blatant lies, just as your misrepresentation of the story.

"if you ask them to show you what it is about AI art they love so much" - blatant lie. The reason this image was linked was your attempt to misrepresent AI art by posting you own lame "creations" as an example of inadequacy of AI art in general. It was accompanied by the artist's description of 6 hour work process to counter your another attempt of misrepresentation - "it's just a prompt".

You can't make a single argument without obvious blatant lies. Maybe you should stop trying because I am not going to ignore it.

 

[Tall Dwarf]

I know this is a tech site—and we often take fairly exotic tech ideas for granted—but it is utterly unhinged that anyone here is trying to argue that learning to prompt an AI is in any way a more accessible method of creating images than a pencil an paper.

Utterly unhinged. Completely detached from any semblance of reality. Like trying to assert that yacht racing is a more accessible sport than soccer.

Once you have a pencil and paper, that's all you need. You can draw. You have accessed drawing. Everything else is optional. The minimum tool for creating AI images is "computing device." That is literally orders of magnitude less accessible.

I went to the KAWS exhibit at the Art Gallery of Ontario a few weeks ago, and one of the most unexpectedly fascinating things to me were some of the pen sketches and the way he creates the idea of heavier lines in them. Sure, he has lots of other media available to him, but he can still get stuff done in interesting ways with materials that people will literally give away.

You forgot the time. Thousands of hours of free time you need to make anything half decent.

Well, you didn't forget really, you decided to pretend you did.

 

[Dail]

Gen AI has been a game change for my illustration practice TBH. Illustration for me takes a long time and being short on that, I drifted away from it. Once I tried out Midjourney and in particular style references and character references which I based of my own prior body of work I've been able to 'draw' / ideate at a much much faster rate. Its been a revelation. The sheer unpredictability of it all is enjoyable. The AI outputs are then just part of a process, taken into Photoshop / illustrator and reworked into screen prints.

I find the whole gen ai is derivative and theft a bit of a beat up. Everyone is derivative, its how visual language evolves. All artists / designers / illustrators stand on the shoulders of giants.

 

[Marcus Andreus]

You forgot the time. Thousands of hours of free time you need to make anything half decent.

Well, you didn't forget really, you decided to pretend you did.

​

(https://xkcd.com/285/)​

 

[daggar]

I fear this just emboldens AI powered competition. The door was unlocked and swung open.
Like it or not - and put me in the "not" category - many people LOVE gen AI because of the ease to flood the market with junk. Big gamble...

The AI powered apps are not competition. They are garbage generators. If someone wants garbage, they can load it in from one of those apps. Procreate is not losing out on anything, AI is just a race to the bottom reliant on venture capital to pay for the shovels. Once

 

[hizonner]

"And it starts to look like a fairly unpleasant form of work" - an opinion. People are doing this stuff in their free time because it's pleasant to them. Your tastes are not universal. Big surprise, I know.

So accessibility (which includes engagement) depends on who you are. As you keep trying to deny.

We are talking about accessibility in US. You are trying to extend it because you are intellectually dishonest.

I am not in the US, and no I was not talking about the US, and I don't live in a tiny world consisting only of the US. Nor did you say anything at all to limit it to the US.

But I believe more people in the US can draw a face than have tried to use AI to generate an image. Even one single image. I know of no actual, non-ass-pulled statistics anywhere. Which is why I don't asspull actual quantifications like "tens of millions" of anything.

Yes, it is obviously more engaging. Otherwise all these people would be sitting with their pens and papers and no one would ever notice them.

It appears that you aren't noticing them, anyway.

Tens of millions of Americans used generative AI.

Citation needed. For images, remember now. And I'm tempted to say more than once and at a deeper level than throwing out a few words about the primary subject, since you feel free to exclude toddlers and even adult beginners because what they produce doesn't reach your technical standards.

You. Don't. Know. That. So. Stop. This. Pretentious. Crap.

Talk to me in even 20 years. I'll probably still be alive.

... and retract your noisy prediction that a random "STEM degree" is a better "investment" for the next 40 or 50 years. As if that were somehow the only issue.

And your daughter already determined that her pen and paper skills are a dead end.

In terms of making a living, yes... for however long it continues to matter. And that was one of the reasons she didn't make it her formal educational concentration.

Yet she continues to draw. Almost as if making a living were a stupid thing to use as your only criterion for what's a "dead end".

 

[andygates]

Careful now, you're close to bursting the bubble. There are a lot of people here desperate to believe that if not for this pesky AI business they could somehow make a fortune doodling on DeviantArt. They don't want to hear anything about what making a living as an artist is actually like, that would ruin the fantasy.

DeviantArt is awash with AI slop and AI bot accounts liking each others slop.

 

[omniron]

Gen AI has been a game change for my illustration practice TBH. Illustration for me takes a long time and being short on that, I drifted away from it. Once I tried out Midjourney and in particular style references and character references which I based of my own prior body of work I've been able to 'draw' / ideate at a much much faster rate. Its been a revelation. The sheer unpredictability of it all is enjoyable. The AI outputs are then just part of a process, taken into Photoshop / illustrator and reworked into screen prints.

I find the whole gen ai is derivative and theft a bit of a beat up. Everyone is derivative, its how visual language evolves. All artists / designers / illustrators stand on the shoulders of giants.

Yep. Ai is just a tool that both novices and experts use. Procreate is going to die from this choice, or change their mind.

Here’s a good example of how it enhances productivity
https://www.reddit.com/r/comfyui/s/jlnMnZ2c0U

 

[Cthel]

Yep. Ai is just a tool that both novices and experts use. Procreate is going to die from this choice, or change their mind.
[snip]

Why?

Are you also predicting that the manufacturers of artist's pencil sets are going to die, because they can't get GenAI into a 4B pencil?

Are Windsor & Newton going to go bankrupt because they can't work out how to run Dall-E on gouache?

 

[HeartbreakOne]

Careful now, you're close to bursting the bubble. There are a lot of people here desperate to believe that if not for this pesky AI business they could somehow make a fortune doodling on DeviantArt. They don't want to hear anything about what making a living as an artist is actually like, that would ruin the fantasy.

It's an aspirational goal to ply a living off of your own craft, yeah, but we all work day jobs and feed that degenerate compulsion to create something ourselves in our own time.

When everything except "value to shareholders" is extinguished as a concern, you'll have your paradise, I imagine? I hope you get it.

 

[ThatEffer]

Wait... What is Shitgasm Art Fucker's stance on AI, anyway?

Given that they fuck art, they're all in!

 

[deathtothenonbelievers]

A operator searching the output space of a generative AI is entirely beholden to the skills of the artists whose inputs were used in the machine learning phase. A prompter can only ever draw out and combine the patterns that are contained within the neural network. The novelty of generative AI is entirely limited to the yet unthought of combinations of elements that it can effectively extrapolate into a coherent image. How am I to regard an individual as an artist, when all they have done is stated their whims to a machine and hoped for the best? The only skill that person has developed is an ability to call upon machine-borne heuristics developed off the backs of countless uncredited individuals. What is there to appreciate beyond the underlying machine and beyond my own pattern recognition?

Art is not some inaccessible height that many will never reach. The primary costs of becoming an artist is time. Although formalized training and fancy tools might help, and accreditation might turn a hobby into a profession, there are plenty of self taught artists, who paid for their expertise with nothing but their free time.

And we can all be physicists if we just put in the work. It's laziness, not lack of talent, that prevents people from doing anything. Same for monkeys.

 

[Penforhire]

I am feeling a good analogy here. When digital cameras proliferated it let anybody and everybody take photos that are generally in focus and likely reasonably exposed. Film cameras got to a similar point of useability but still had a 'cost of entry' to develop film (and prints or scans because nobody likes to look at negatives as end products) or transparencies. But there is a vast chasm between that and being a skilled photographer (composition, timing, precisely intended exposure and lighting, point of view, post processing, etcetera), even aside from whether the result is "art" or not (a meta or emotional judgement).

Feels similar to arguing these AI-lite models can create art because now anyone who can throw a description together can generate an image.

 

[hel1kx]

And it was actually awesome:

In your opinion maybe.

I see tons of folds in clothing that don't make sense, the dragon's mouth somehow being on two different depth planes/coming out of a rock that somehow is also the dragon's body, dust/smoke appearing out of clothing nonsensically, hair moving in a totally unnatural manner, a random wing/rock coming out behind the person, random scaly artifacts that aren't really attached to anything.

It might look cool to you, but to me it looks unnatural. Yes fantasy stuff can be unnatural, but at least make sense.

Now, the client wants the girl's head to be turned to face the viewer, but everything else stay the same. Good luck doing that with a generative AI program.

 

[JoHBE]

"Externalised" as in "it allows new people to compete in the market". Which is not quite as tragic as global warming.

"But will end up dissolving the very fundaments it is standing on" - or not. The whole thing is 2 years old. It is quite possible that training doesn't need billions of images to succeed, this is just how the current generation of tools work.

As we speak, it's outright destroying several online communities by out-flooding Human output and/or completely wiping out and obfuscating the distinction between what the Human put in, and what was hallucinated by AI. This goes for writing, too, btw, referring to the online sci-Fi story submission site that had to close down because moderation is just impossible to scale.

Fun.

 

[iindigo]

I find the whole gen ai is derivative and theft a bit of a beat up. Everyone is derivative, its how visual language evolves. All artists / designers / illustrators stand on the shoulders of giants.

Generative AI is derivative to an extent that no artist other than maybe painting counterfeiters are. A human artist doesn't need vast libraries of existing work to operate. In fact we can produce work and invent new styles by simply putting to paper what we imagine. Generative AI on the other hand is like a play-doh press: if you have no play-doh (art) to put into it, it can't function.

It's not hard to imagine future versions of generative AI that don't need such extensive training and can produce something intelligible with a very small dataset, and when that happens it'll be about as derivative as humans are, but there's a long way to go before achieving that.

 

[JoHBE]

Your pet theory that AI will eventually collapse on itself is just that, a theory. AI generation has the only way, up. Because if a new model is inferior it is scrapped.

That will also happen, but Imeant that it will absolutely destroy the value of anything it touches, since it turns everything it touches into commodity.

 

[iindigo]

In your opinion maybe.

I see tons of folds in clothing that don't make sense, the dragon's mouth somehow being on two different depth planes/coming out of a rock that somehow is also the dragon's body, dust/smoke appearing out of clothing nonsensically, hair moving in a totally unnatural manner, a random wing/rock coming out behind the person, random scaly artifacts that aren't really attached to anything.

It might look cool to you, but to me it looks unnatural. Yes fantasy stuff can be unnatural, but at least make sense.

Now, the client wants the girl's head to be turned to face the viewer, but everything else stay the same. Good luck doing that with a generative AI program.

Its composition also needs a lot of work. There's no path for the eyes to follow, no usage of color, shading, or line weight to distinguish or emphasize the important parts of the image. It lacks intent, with the whole thing grasping for the viewers' attention all at once. Its entire value is tied up in being glossy looking.

Of course, these are all mistakes that human artists can and do make, particularly when starting out, but none of those artists hold up their work as outstanding or masterpieces. On the contrary, human artists are very frequently self-critical to a fault to the point that it's not unusual for masters with decades of experience to still doubt their work. The work of self-development never ends.

 

[Aurich]

In your opinion maybe.

I see tons of folds in clothing that don't make sense, the dragon's mouth somehow being on two different depth planes/coming out of a rock that somehow is also the dragon's body, dust/smoke appearing out of clothing nonsensically, hair moving in a totally unnatural manner, a random wing/rock coming out behind the person, random scaly artifacts that aren't really attached to anything.

It might look cool to you, but to me it looks unnatural. Yes fantasy stuff can be unnatural, but at least make sense.

Now, the client wants the girl's head to be turned to face the viewer, but everything else stay the same. Good luck doing that with a generative AI program.

I make fun of it because after all this high minded discussion, all this "look at the choicest of prompts, this took 8 hours of organic free range prompting" the end result is just the most banal fantasy art.

It's a vague rip off of a bunch of people's styles, none of whom had a particularly original style to begin with. It's a book cover for a self-published fantasy novel series you skim past on Kindle Unlimited.



I'm mocking it not because I hate it, not because I have any ill will towards the college student who spent an evening having fun making it, but because all this "your degree is now shit and nobody should ever learn to draw again" championing is over ... this.

It's a race to the bottom, and the only winners are people who want more noise. Great, content is cheaper and faster. Is that really what the problem was?

 

[Cthel]

[snip] On the contrary, human artists are very frequently self-critical to a fault to the point that it's not unusual for masters with decades of experience to still doubt their work. The work of self-development never ends.

This is why all the bleating about "gatekeeping artists" is so ridiculous - every artist knows what it's like to fail to hit the platonic ideal that only you can see*, so no artist is going to tell someone that their art is "so bad you'll never be an artist"

The only people who'll say that are people who've never tried to create something.

*or not, as the case may be. Being an artist with aphantasia can be infuriating

 

[S2pidiT]

You claimed "you don't get to skip" and it is objectively wrong - you absolutely do. The new tools allow absolute beginner generate awesome images. With some skills and efforts they can exercise remarkable control over the process and the results.

So... more micromanagement tools available to middle and upper managers, except they don't even need workers now!

You changed "creating art" to "generating awesome images," which are not the same thing. As I said, creating art is the journey, and skipping means you're literally not taking the journey. It makes you the quest giver, not the adventurer. The commissioner, not the painter.


But I don't expect my part of this discussion to change others' minds, so that's enough for me.

 

[EnPeaSea]

It's a book cover for a self-published fantasy novel series you skim past on Kindle Unlimited.

View attachment 88443

YOU MONSTER!

 

[s73v3r]

"I want this person's arm here, not there, and the background to be composed like this, and the shadows to do that" - it is possible already, for quite some time. It's called ControlNet. And the new models are very good with prompt adherence so stuff gets easier all the time.

"from where she happens to be sitting right this second, she could stick out her arm and pick up a pencil and paper" - you do understand that this personal anecdote doesn't change accessibility of art for an average person? She is not typical.

And even for her - the skill she has has no market value. She can go learning art, get in debt and obtains a useless degree. Or she can get superior results by just learning AI tools in her free time while studying STEM.

The person didn't do a shred of work. They had the computer do it all for them.

And you've still proven the entire reason you're on this bandwagon is because you're insanely jealous of creatives that can actually make art.