Any artist jumping on this now Adobe provides this tech, tech that has been around for months now, for free. Deserves what's coming for them.
I PROMISE you once Adobe has market saturation on this and people have embedded it in their workflows and work without it feels painful they're going to be slapping a $40-$100 (It replaces literal days of work and literal team members, they can charge what they want) a month price tag on it on top of Creative Cloud.
Artists once again, never learning from the past and handing over their future to an abusive company that seeks rent on their ability to earn.
All you had to do is install any of the countless Stable Diffusion UIs and you'd have this tech free forever, some of them even integrate into Photoshop.
No sympathy for anyone who chooses this path when we finally had an opening to harm the Adobe monopoly.
Yep it's easy... just download a wrapper app from GitHub, download the models, set up the configuration, realize you're missing libraries, install a specific version of each etc
Your average user doesn't understand how to set up and leverage open source software because it's too complicated to get going.
They want to open Photoshop and use the feature in the tool the already use, without setting up anything.
No offense but you sort of prove the parent’s point. People who work on computers professionally but don’t use open source because “it’s too complicated to get going” are making an error. Adobe’s UX being addictive is the point.
Many products have been successful not because they bring new features to the market but because they make existing features trivial to use. UX is everything.
It really is. Arguably the most famous example among HN readers was BrandonM’s comment[0] on Dropbox’s launch thread. Was he correct that you can built it trivially using 2007 software? Sure, if you had the correct mix of specialized knowledge of the software and networking. Dropbox was not a novel idea, but that’s now where it value was; the value proposition relied 100% on its UX.
If MSFT and Apple each shipped a Dropbox competitor that was highly integrated, available, and pushed flagship software suits to adopt it as well within a year, would Dropbox had become as large as it did? I would guess not. It would likely be a downgrade from what Dropbox offered (the separate ecosystems staying bifurcated comes to mind), but a zero install solution would be the only thing that could make it easier.
No offense taken, I'm simply pointing out that the
user experience of setting up and leveraging open source is a challenge to wider adoption.
I expect most of us (hackers) have tried these tools, but not your average user. When I see comments like OP, it's coming from the hacker perspective imo.
The user is not making an error, the error is assuming everyone should be a hacker.
The user is definitively making an error here. Especially when the user is actually not a user but a producer (supposedly artists). As an artist you should strive to have a certain level of ownership and independence.
It gets worse, there was a thread a couple weeks ago about Envato (ex ThemeForest) asking all their authors for "free/unlimited" rights for anything they have to be used to train AI models for NO compensation. It was eye opening how many "artists" and "producers" accepts the status quo as is, and think it's still a good idea to collect whatever this company throw on their way.
It's really hard to feel sorry for these people...
How many work hours would it take for the average professional artist to get to this level of technical proficiency? Given the opportunity cost I'm sure you could argue it would be a net loss for the employer.
Employers don't need their graphic artists to be able to use git and navigate Python dependencies. That's why they have IT.
The mistake companies make is in not fully embracing this.
Alright, once you get it running the UI for generative fill can be used by a graphic artist, and now you're several months ahead of your competitors, but still the UI is a little rough. But it's open source, so you have your techs make it a little better.
Now your company needs to make it easy to submit a pull request, or you're suddenly maintaining your own fork, which you're going to want out of as soon as you have an alternative, and soon your company is paying $$$$ to Adobe again.
I work with a lot of people that don’t care. They have a job with deadline, for which they already gained a lot of knowledge, and need to keep it updated. So no, the supplementary friction is not acceptable, even for « good » reason.
What type of "work on computers professionally" are we talking about? Working on a computer as a professional artist or writer is much different than working on a computer as a programmer/hacker professionally. For the first, the computer is just a tool like a pencil or a paintbrush. I wouldn't expect them to know how to figure out how to use things unless it has a very very simple UX.
its not addictive, its cutting edge and has a price tag for that. open source will follow up with the cheap entry point for the nerds. If you dont like that, then you need to throw proper money at your favourite github accounts because that is fundamental difference. Adobe is funding high end tech solutions by selling a product. if you have another way to to achieve that, do tell.
too many college canteen marxists in this place is my thinking.
From an economic standpoint, this makes sense. If an artist is bringing in 4k a month, then spending 40 a month for software maintenance and updates is rounding error. This also opens up the third part market to sell plugins if open source doesn’t work for them. All in all this is win win for everyone and more options equates to more freedom.
That is such a strange thing to read. I'm not contradicting that it's true in this case, but Basically every artist complains about how much of a clusterfuck the UI/UX is while being entirely happy that the UI is horrible enough that intimate knowledge of it keeps them with a job
I'm not saying it's easy, it's Python so its utter hell at times but I'm saying it's worth the effort to have something free forever and easier UIs to it are coming.
Agree with you there, if these tools become 1 click install and go they could get wider adoption. However in the B2B space, they also need parity with commercial tools like Photoshop, Figma etc. It's an uphill battle for an open source project with no profits going up against a profit driven big tech company.
If you're a professional artist who's invested years of effort in learning photoshop and it replaces "literal days of work" or you run company and it replaces "literal team members" $40-$100 a month is cheap.
If not you can use something cheaper - or use one of the 3rd party plugins. If the workflows become popular, the 3rd party plugins will get better.
> If you're a professional artist who's invested years of effort in learning photoshop and it replaces "literal days of work" or you run company and it replaces "literal team members" $40-$100 a month is cheap.
No kidding. I have only negative things to say about Adobe, but that is a terrible argument. $100 a month to replace multiple $100k/year employees? That's basically the definition of technological advancement freeing labor to be more productive elsewhere.
the definition of progress. In IT we should be used to this. AI has been telegraphing its arrival for a long time. There should be no surprises to see jobs being lost in the name of tech advancement. Its been going on since forever. It is literally what computers do.
The problem with the AI generated images is that it's abundant and anything abundant doesn't have a market value.
So, if an artist's work makes 100$ per work an AI system which can generate the same quality work but 1,000,000 pieces a day this won't end up the AI system generating 100 million dollars a day.
It will simply rise the bar for artists instead. Which is a good thing for the society as a whole because the baseline will be very good and the work of the talented artists will be highly valued.
Adobe can eat the market of mediocre artists if they are the only one with this technology so they can control the output rate and create artificial scarcity. They are not the only one with this technology.
The problem with AI generated images, in this context, is that they can be used to imitate the style of specific artists, meaning any unique talent and vision those artists have no longer has market value. The scenario you describe isn't actually a business model anyone is following, except maybe morons shitting AI porn onto instagram and trying to flog their patreons.
Obviously AI generated art has market value to companies with creative staff because AI is designed to devalue and commoditize the work of actual artists, such that the work of all artists can be replicated by someone on Mechanical Turk writing prompts for pennies a day (or some equivelent low-wage plebian drone.)
AI has already been used to steal commissions from working artists, and companies are already firing their entire creative teams to lean entirely into AI generation. The better the technology gets, the worse the situation gets for working artists because that's the plan.
I don't see how anyone can have witnessed the glee with which AI techbros danced on the graves of the art industry, ridiculing and harassing artists, calling them gatekeepers and fascists, and telling them their careers were doomed, and think this will benefit artists in any way.
I really want that to change, because I find this technology fascinating, but as of now the well has been poisoned, the bridges burned, war declared, and most artists with actual talent would rather starve in the street than touch anything AI.
Yeah, commission-based digital art seems likely to mostly go away, in the same way commission-based oil portraits mostly went away with the diffusion of photography.
It seems like generative AI will wipe out many design shops and non-IC/exec brokers/managers of creatives (recruiters, talent agents, admins, etc.) at least to the degree the web wiped out newspapers and their non-creative staff.
But it'll also enable creative designers/artists/writers to become their own studios. As Kubrick said, "one [person] writes a novel, one [person] writes a symphony." Generative AI will enable one person (or a very small team) to create a blockbuster movie or AAA videogame.
And we already got that with print for a while. You can create an Andy Warhol piece as well as his assistant for a long while now and no the market didn’t collapse or art disappeared. It open some door, close some other, offered new tool.
As an artist (creator) myself, I do not feel threatened. I do not consider the tedium to add anything of value to my work… and hell, it’s suck to integrate /generate fill (as an example)
> The problem with AI generated images, in this context, is that they can be used to imitate the style of specific artists, meaning any unique talent and vision those artists have no longer has market value.
AI can generate an image "in the style of Banksy" but so can a thousand other artists who aren't Banksy, and their work product would have no more value than the AI's.
The way this makes you feel is about more than texture and shading:
You might be able to get Stable Diffusion to generate something like that with a sufficiently detailed prompt, but then its uniqueness would come from the creative effort required in devising the prompt. It's not like you can just type "street art in the style of Banksy" and expect the output to make you feel something like that.
>I don't see how anyone can have witnessed the glee with which AI techbros danced... most artists with actual talent would rather starve in the street than touch anything AI.
Perhaps. But what would that change and do you have any choice? 'Techbros' can make do with genAI or even stock images. I'm not so sure about said artists.
>The problem with AI generated images, in this context, is that they can be used to imitate the style of specific artists, meaning any unique talent and vision those artists have no longer has market value
I Agree, those who produce the training data or the thing that the AIs work will be based on, should be properly compensated.
The artists that are hurt by this have neither a name nor a distinctive style, they are the ones whose work was derivative itself, who already struggled to get paid for expensive manual labor.
Then you have the artists who matter, whose work is valued because it has their name on it, regardless of what process was used to create it. AI will be just another tool they may choose to use.
The narrative that the only artists affected by AI are mediocre and derivative and thus deserve what they get, is just propaganda. AI was trained to replicate the distinctive style of artists who "have a name" and "matter," and those artists likely work for companies that are training in-house AIs on their own IPs as we speak.
"struggled to get paid for expensive manual labor?" You seem to have a contempt for most working artists, because most struggle to get paid regardless of talent, not because they're incompetent but because art of any kind has always been a difficult market. And yes, it's hard work, which many people fail to appreciate. You're really illustrating my point more than I think you intend to.
I'm not saying they "deserve" what they get, not any more so than the people who found themselves out of work after weaving or sewing machines were invented. That's just technological progress.
Perhaps we have a very different understanding of what "art that matters" is, certainly I'm not thinking of anyone producing art at volume for a corporate IP. In any event, nothing ever stopped such a corporation from reproducing a certain style with cheaper labor, whether it's human or not. If you were, say, an art director working for a corporation, your job was never to perform much of that manual labor. AI isn't replacing you.
Moreover, the emergence of mass manufacturing always comes with a counter-movement attempting to maintain the "human touch" and charging a premium for it. For every Budweiser, there's a thousand craft beers. Neither replaces the other.
Either AI created graphics arrived to late for the NFT craze, or they arrived in time to give them a second spring.
Being serious so, human art is what it is because it requires effort. AI art doesn't require nearly as much, is generic and abundant. In short, boring.
There's so many places where you don't need "real" human touched art but you could do with stylized custom images. Midjourney and friends make it a million times harder to make a living doing small time commissioned art, stuff like dungeons and dragons character and world art, furry commissions and smut, and hobby art in general.
I think the incoming popularity of generative AI won’t simply raise the bar, but at first devalue itself, then make AI generations irrecognizable. The public will start to fail to recognize an art is there, when human discretion is not found.
There is a sci-fi prediction that humans would eventually reduce to a thing with an eye and a finger that only presses a button, and machines would do the rest; I take it as an optimistic view that it means humans remain relevant, and that human values only come from humans.
> All you had to do is install any of the countless Stable Diffusion UIs and you'd have this tech free forever, some of them even integrate into Photoshop.
But Photoshop's generative fill is trained on Adobe's own stock photos, which removes an entire class of potential future litigation issues.
The more we perpetuate free and open models the less possible it will be to litigate against them. They can't sue us all.
If we move to this lie that Adobes models are "ethical" it gives them a moat to make free models illegal so Adobe becomes the only option.
Artists should be realizing the goal is to stop Adobe becoming the only option, if they do become the only legal way to do this then expect AutoCAD level pricing.
> Artists once again, never learning from the past and handing over their future to an abusive company that seeks rent on their ability to earn.
I think this is a little bit unfair and trivialises the problem. I know many artists who have to/had to use CC for their work because their clients or (mostly) employers consider it the industry standard.
I agree with the sentiment, fuck Adobe and decades of their parasitic practices, but the reality is not that black and white.
If you already pay for Adobe CC as many designers do, there's an opportunity cost to learning an entirely new system and workflow. Those are hours that can't be billed. The opportunity cost of switching would make sense if Adobe was charging 5x more for these features, but they're not.
At the end of the day, not everyone will need to "make" generative art. There will be plenty of it on Adobe Stock that clients will be fine with, especially once they see the hourly billing rates for custom generative art that requires multiple prompts to get right.
> All you had to do is install any of the countless Stable Diffusion UIs and you'd have this tech free forever, some of them even integrate into Photoshop.
You mean the ones trained off the artists art without permission, unlike Adobe's offering?
I know a lot of graphic designers that looked at that and hated it because of the fact that they took stuff without permission.
I don't know a single artist who was happy with that. The AI generation wasn't the issue. It was the training source.
I'm not an artist, and I'm also aware of the many failings of Adobe, but you really missed what the issue was.
The alternative is more likely - it becomes commoditized and many programs are offering it, so it would be strange for Adobe charge a premium for a baseline feature
>so it would be strange for Adobe charge a premium for a baseline feature
Alternatives to every baseline feature they currently charge a premium for exist, but because its industry standard everyone has to pay them.
Nothing Photoshop does today is special, in fact some of the ways it does things are worse than everyone else, e.g many filters being single core constrained and CPU bound.
You seem to be falling into a fallacy common among technical people: that a product's only value is its technical implementation.
Photoshop is sold to businesses. It's value prop is "you will be able to hire anyone and they will be productive day one, which will save a lot of money in training."
People buying photos op don't care about filters being CPU bound. They care about turning work around quickly and getting the next paying gig. A CPU bound filter is statistical noise compared to having to figure out how some different tool works.
> You seem to be falling into a fallacy common among technical people
I work as a creative director, my anger about this is decades of suffering what I consider substandard tools.
> People buying photos op don't care about filters being CPU bound. They care about turning work around quickly
We definitely do care about our computing power being used to the fullest of it's potential. This is why some of us even do 2D static work in After Effects now because its faster and uses more of the machines power than PS.
Photoshop developer asking: what examples can you give of 2D/static work that performs better in AE over Ps? I'm very curious about your workflow now...
Well I apologize for my presumption. Showing a little of my own frustration as a former engineer, now product person, I suppose.
I'm really curious why you care about using machines to their potential. Is it cost savings from not having to buy a more expensive machine, philosophical dislike or waste, or does it really come down to hours of non-billable time?
Adobe’s UI is also terrible, but it’s terrible in a way most digital artists are familiar with. Starting from scratch Gimp, Inkscape, and Adobe have reasonably similar learning curves.
The main advantage Adobe UI has is many people assume they’re stuck learning it at some point. But it’s definitely a waste of time and money if you’re happy as a hobbyist.
Adobe doesn't count hobbyists as a core user segment; why would they? Hobbyists don't need Adobe CC any more than somebody looking to multiply 2 numbers would need Excel 365.
The hobbyist market is already served by Photopea, Canva, Snapseed and many others.
Adobe has figured out how to charge a lot of Photoshop and likely will do the same for major additions to it. You can argue whether this is fair or right or that there are free atlerantives, but it is possible for Adobe and so they will do it.
So the fundamental problem and “promise” is that a company charges for the products it builds, or that professionals “deserve” to be paying for the tools that enable their craft?
You don’t have to pay for things! The models are just there! The code is just there! The images to train on are just there! The absurdity that you have some rights over some data is the only thing that could possibly cause you to pay money for a thing.
There's no free lunch. There's a reason the first law of thermodynamics is first.
Atoms are just there for grabs too. Yet, you pay for someone organizing them in some order.
At least that's how the world currently works. You can walk for "free" or take the bus for a "fee".
So nothing is actually "free", although the assumption that Adobe took some off the shelf models and off the shelf pictures and threw them in Photoshop, didn't take much effor either.
Creative Cloud is less specialized now, not more. They were able to defend their margins because designers were used to specific workflows. As the interfaces become more general, that advantage will erode. Adobe is working on sophisticated chain of custody for DAM which won’t matter to the bill of CC users, even in enterprise.
The productivity gains are becoming permission-to-play. The money saved will go into hiring different kinds of people which will continue to pressure the software budget. It’s true CC is seen as necessary, but it’s still resented and there are always employees whose access to a subscription is marginal, providing some elasticity.
>I PROMISE you once Adobe has market saturation on this and people have embedded it in their workflows and work without it feels painful they're going to be slapping a $40-$100 (It replaces literal days of work and literal team members, they can charge what they want) a month price tag on it on top of Creative Cloud.
I think 100 a month is a low price for this.... if it saves literally man-days of work. The equivalent value addition is something in the ballpark of 3k per project (assume 100 an hour over 3 8 hour days).
Add in the fact that your velocity has increased significantly, they could probably get away with at 5k a month at the "Pro Tier"
> All you had to do is install any of the countless Stable Diffusion UIs and you'd have this tech free forever, some of them even integrate into Photoshop.
this is the "anarchist" hacker mindset; fear of "the man". open source is always going to lag behind the huge, well funded corporate option because they have teams of highly paid tech working on being first out the traps. open source and corporate dont really compete except in the minds of "anarchist" hackers who don't like capitalism. Yet both worlds can exist, and do. Adobe have produced a mind blowing product. Open source versions wont be far behind. Both pathways exist in parallel. Open Source wont be funding the top of the range stuff ever.
I PROMISE you once Adobe has market saturation on this and people have embedded it in their workflows and work without it feels painful they're going to be slapping a $40-$100 (It replaces literal days of work and literal team members, they can charge what they want) a month price tag on it on top of Creative Cloud.
Artists once again, never learning from the past and handing over their future to an abusive company that seeks rent on their ability to earn.
All you had to do is install any of the countless Stable Diffusion UIs and you'd have this tech free forever, some of them even integrate into Photoshop.
No sympathy for anyone who chooses this path when we finally had an opening to harm the Adobe monopoly.