[nkrisc]

It’s sad how the existing corpus of artists’ works are being used against them.

The only reason this can generate images like these is because artists previously created artwork like this. And then in return they get:

> you don't need to high an artist or animator to develop you game.

Unless you created all the training material yourself, any model like this is highly unethical, in my opinion.

[archerx]

As someone who is an artist and also a programmer I find the differences between both very different. I never see programmers crying over A.I. chiming “they stole our code so it’s evil and I’ll never use it and neither should you!” Like the artist seem to do.

I don’t really care if the image gen models are trained on my renders or photos as long as they are open source.

Also I have been making the assets for my game which needs at least 2,000 sprites and honestly it’s very tedious and I’m looking to automate as much of the pipeline as I can so I welcome anything that removes the the tedium and pain from the process.

[asddubs]

people did complain about it stealing GPL code scraped off github etc and then license laundering a lot, especially during the earlier days. I think now people are just worn down to have the conversation again.

I do also think the threat is bigger to artists. Sloppy code will cause problems, and at some point you'll need someone who knows what's going on to step in. Sloppy art, a lot of people accept.

[bandoti]

I’m not sure over time people will continue to accept sloppy art.

The “glitchiness” of the AI sprites can be quite unsettling because one can feel the something-quite-wrong that is difficult to place.

This is a worse state than using even rough illustrations by actual artists.

If one must use AI for this I would recommend having an artist in the pipeline who uses AI to create assets but then makes sure they are seamless.

[archerx]

That's why artist should incorporate gen ai into their workflows. They have the artistic eye and separate the slop from the gold better than non artistic people. Also knowing composition and etc helps getting better results from Stable Diffusion, Flux and etc.

I remember when 3D CG was new and it got so much hate for not being "real art". I heard the same drama happened when Photoshop and Illustrator hit the market, they weren't considered "real art".

[asddubs]

artists like making art and having some sense of intentionality, rather than just typing stuff in and then messing with it. that's the difference between cg and ai. you cannot cut intentionality out of the process and then just do tweaks to have it not be slop, at that point you cease saving time.

[davidclark]

From a recent blog shared on HN, among a list of reasons they can’t or won’t use LLMs:

>The training data for LLMs is stolen. I don’t mean like “pirated” in the sense where someone illicitly shares a copy they obtained legitimately; I mean their scrapers are ignoring both norms and laws to obtain copies under false pretenses, destroying other people’s infrastructure. [footnotes omitted]

“I think I'm done thinking about GenAI for now” https://news.ycombinator.com/item?id=44193018

[archerx]

I do agree that the scraping is annoying and I had to set up some anti scraping measures on one of my image heavy sites. However I am for the freedom of data especially if the models are open source. I only use local models and haven't used Claude or ChatGPT since last year and it's pretty awesome.

[davidweatherall]

I think there's a big difference with how designers and programmers can use AI to enhance their work, and how exploited the average person's work has been. A lot of designers publish their work on social media, or it's visible on their websites etc., which AI models have used to train with. But with coding, most people's code only exists in private repos, or is compiled to a format that LLMs cant easily be trained on.

[archerx]

There is so much code on github and all the other repos that it can compare with the amount of art that is out there. Also all of your Javascript code is public once your site is published.

There is nothing stopping artists from using A.I. to improve their work. The sketch to image work flows are great, the variation work flows can save from the tedium and inpainting can help fix and improve images.

Text to image is lazy and is the cause of most of the slop but no one is saying artist should replace their Wacom tablets with text prompts. I feel like there is a lot of hurt egos going on. I remember back in the day on CGtalk, there was so much elitism and in hindsight it probably held a lot of people back, myself included.

[nkrisc]

I’m not opposed to generative AI, in principle, it’s how it’s being employed that bothers me.

[archerx]

Please elaborate.

[nkrisc]

It bothers me when it's sold as a tool to replace human creativity. I do believe it can absolutely be a tool that aids human creativity. I would rather see more of the latter than the former.

They say, "you don't need an artist with our AI" instead of "you too can learn to become an artist with our AI". The vision they're selling is a world without artists. They want the art without the arist.

[Aeolun]

Unless you didn’t want people to see and use it, maybe you shouldn’t have made it public? I don’t think training AI is fundamentally different from humans learning from things they see, and we don’t restrict that either.

[wrasee]

The fundamental difference is that computers can do this at a pace and scale that humans could never aspire to. There is a natural limit to the extent a single individual can be informed by previous works. It sets a natural pace to innovation that is sustainable for both the artist and derivative works.

Computers have no such limitation and can consume almost the entirety of a subject’s work in a few weeks or months. I think that alone is enough to say that yes, it is fundamentally different.

[jason_oster]

Automated assembly lines have the same properties. Same with transportation. Buses, trains, airplanes, ships. These all work tirelessly at a pace that humans cannot match.

There is no fundamental difference.

[wrasee]

I think you missed the point. The claim is that AI training from public information is no different from humans learning from public information.

My argument is precisely that the mechanisation of information is fundamentally different from the scale at which a human can learn. One immediate consequence being there is no longer a natural brake on the scale of what can be sourced for use in a derivative work.

To be clear this is not a value judgement, just to point out that it _is_ different, just as driving is fundamentally different from what one can do with one’s own feet. Of course the mechanisation of transport is history and seems daft to argue against. But it is different. Whether that’s good or bad is a much harder question.

[Aeolun]

I can agree that it’s different, but driving is not fundamentally different from walking, in that both get you from one place to another. Nobody drives to a place because it’s fundamentally different from walking there, they do it purely because it’s faster and leaves them less tired.

I think the same thing is true for AI. Or at least, for training or acting on public information. It’s not suddenly bad because you are able to do it on all information in existence at the same time.

[fourside]

When it’s big copyright holders we have very specific, very granular definitions of what constitutes fair or allowed use. But when it comes to smaller creators the answer is that it’s their fault for trying to promote their work and make a living.

[Aeolun]

No. It’s not fundamentally different. Big copyright holders are better at asserting their rights, and had their information under lock and key from the start, but I don’t think for a moment all these models haven’t been trained on all the content on the disney website (or all the websites with disney derative work).

Disney is just better at preventing the LLM companies from allowing it to regurgitate that stuff.

[bnop]

1. These models have been trained on private data many a time without permission

2. Making something public isn’t always a choice made by the creator

3. Making something public does not denote fair use, this is why copyright (albeit arguably a poor solution) exists

4. These LLMs consolidate wealth into a small group using outputs from a larger, often less wealthy group (creatives) without fair compensation

[Aeolun]

1. I’m disinclined to believe that purely because of the inconvenience of doing so. Much easier to scrape the entire internet.

2. How so? If you sell your stuff and someone makes it public, it’s still your choice to sell it.

3. That’s only true as far as recreating the content is concerned. Reading and viewing is by definition fair use for publicly visible information.

4. Define fair compensation. I feel like the creatives are just upset that their work is “easy” to replicate with these models. And that isn’t even true, they never appear as unique or interesting as truly new work.

[bnop]

1. If scraping the entire internet is easier, but doesn’t give me as good of results as including private or licensed data, I will train my LLM on private data or lose the race. It is not about easy. Just look at some of the lawsuits against OpenAI: https://originality.ai/blog/openai-chatgpt-lawsuit-list

2. Think about this - what you are saying is that you can’t sell anything without also making it public. These are not the same thing. I sell something to get value from my labor. I make something public to get eyes on it. The whole issue is that people who want to privately sell things have their work being undercut by LLMs

3. LLMs are recreating the content

4. Take Miyazaki. He spends his whole life developing a unique art style and skill. Years. He makes his living and provides a living for others with it. The value of that _used to be_ the movies he was paid to make and the revenue they generated. Now LLMs can create his work for free, and he doesn’t see a dime whenever someone converts their profile picture into his style. This is the ethical problem - he is not compensated, let alone the ethics of upending artists years of work for the sake of it

[dr_dshiv]

These LLMs primarily distribute intellectual and creative wealth from media conglomerates to anyone on earth with $20. (Without fair compensation, agreed)

[nkrisc]

AI isn’t human.

[Lerc]

This is true, but not an argument in itself.

AI is also not American, A puppy, or a cardboard box full of fairies.

Your underlying presumption is that there is some property of Humans that makes the distinction important. Without identifying that property it impossible to evaluate the merit of any claim about it.

[nkrisc]

I was replying to this, specifically:

> I don’t think training AI is fundamentally different from humans learning from things they see, and we don’t restrict that either.

It places humans and AI on equal footing, which I fundamentally object to. No, we do not restrict how humans learn, nor do I believe we should. I do believe we can, and ought, to have restrictions on how technology is used within human society. Those restrictions may change over time and adapt, but I wholeheartedly disagree with the premise that AI learning and human learning is not fundamentally different - it is different because one involves a human, whose needs should be placed above those of a machine.

It's the tendency to equate humans and AI that I find both distasteful and potentially dangerous.

[Aeolun]

> It's the tendency to equate humans and AI that I find both distasteful and potentially dangerous.

In the sense that they both consume information to learn, they seem remarkably similar.

[brookst]

Can you point me to these artists who created new work out of nothing, without learning from other art themselves? Because every artist I know has worked hard and learned from what’s gone before.

[bayarearefugee]

Even as someone who leans generally anti-corp I would love to see Capcom come in and sue over this since there's no way to look at those demo animations and not see the "training" was entirely lifted from the Street Fighter series.

[akritrime]

Imagine if writers complained that stuff they wrote were being used in literature classes that teaches people how to read and write because how sentences they once toiled over for hours would now be used in mundane letters.

AI can't take away an artists creativity, just like how photography didn't kill painting. Yes, it closes up certain avenues on how artists make money but it will most probably make up for that by making them more productive.

[Draiken]

Comparing teaching other humans to "teaching" a model is absurd.

AI can't take their creativity but it will take their means of survival so a few companies can profit instead. Yeah, great trade-off.

[Lerc]

>Comparing teaching other humans to "teaching" a model is absurd.

Bold conclusion, but why is it absurd?

New ways of making thinks has always had an impact on the old ways of making things. Are you proposing we should seek no new ways of making things? I have been described as a creative person by some. The notion of restricting what people can make and how they can make them does not gel with my idea of creativity.

I am sympathetic to those who might have their source of income affected. That is not a problem with the current change, it has been a problem with all change throughout history. If the welfare of people is really your concern then your issue is not with the change but how society supports people effected by that change.

[Draiken]

> Bold conclusion, but why is it absurd?

It doesn't learn anything, it simply regurgitates words based on probability. How we can compare a human learning from other humans to a machine mathematically accessing a zip file with human knowledge is beyond me. It's simply not even in the same ballpark.

> New ways of making thinks has always had an impact on the old ways of making things. Are you proposing we should seek no new ways of making things?

Let's not pretend that any of this is about making things easier or better for everyone. This is to make money and any benefit is a mere side-effect.

[akritrime]

The point isn't about comparing teaching and training. The point is about how humans use tool to improve and advance. It is more about comparing AI to books as an aggregate of knowledge. Just like how humans uses recorded writing in the form of books to analyze and adapt better ways of communication, humans can also use AI to be productive in fields that was previous out of their reach.

[user____name]

> Yes, it closes up certain avenues on how artists make money but it will most probably make up for that by making them more productive.

How?

[akritrime]

In a few years as image generation consistency improves, any video game artist can go from just creating characters for some other game dev to creating full-on animated sprites directly. Instead of having to depend on someone else to code their ideas, they can directly create that game. And the inverse will also be true, game devs can use AI to create generic arts for their game and work faster, create games that depends more on tech than art. Both flavors of games will bring something unique and doesn't invalidate the work of the other as useless.

[snackbroken]

It's already happening in the movie industry. Things like painstakingly hand-tracking an actor's face onto a body double vs using deepfake face replacement. A single vfx artist can now do this work much more quickly, pushing down the cost to do it. This in turn opens up new opportunities for work which were previously not economically viable, like splicing in retakes or making mouth movements match the voiceover dialogue.