I say this as someone who frequently uses generative ai, and actively chooses to pay for the service.
Fuck openai.
This company has utterly failed to fulfill their mission statement, and they will be unable to make right by humanity until ALL software they have created is available to the public as FOSS (free and open source software). Openai claimed that this is exactly what they were going to do, and then they just didn’t. So fuckem.
Ai is like a tool from the future given early to a society of unevolved people. It doesn’t fit the structure of our civilization yet. Until human beings unfuck their animalistic selves it is going to be negative.
If there was universal income, and people didn’t need to work to survive, then Ai would work with society and peoples ideas would grow at a fast rate excelling humanity’s manual creation. Kind of like China’s IP laws and the growth of tech due to the ability to use other people’s creations to build upon.
Also this reminds me of hip-hop and sampling other musicians music.
What is this article even talking about? It’s making no sense.
They’re trying to make some type of argument that a private studio should have exclusive rights to a specific style of art and that by openai allowing users to generate art in that style, we are slipping into anti-democratic authoritarianism.
My opinion is that you can’t own “styles” of art and that there’s nothing wrong here. Legally speaking I can copy any art style I want.
Thanks for that explainer. I thought the verbiage in the article was a little over the top.
However there is a point at which the “style” of the art is the thing that is copyrightable, sort of by implication.
The standard for proving a copyright violation where a defendant claims a transformative use or a derivative work is “substantial similar.”
For as long as I can remember that includes the overall presentation of the work, and it’s hard to describe that as anything other than a “style.”
The article draws a comparison that allowing copyright protection for styles would be like allowing copyrights for entire genres. I don’t think that’s right. Nobody could copyright all “landscape paintings” as a genre, but look at landscape works by Katsushika Hokusai, and that style, to me, is creative enough to warrant protection, if it were made originally in America today and not already in the public domain. And he didn’t invent woodblock prints or even woodblock prints of landscapes, but the way he did it is so unique as to be insperable from the copyrighted work itself and arguably deserving of protection simply for its advancement of the art.
If you made a woodblock print in the same style but used it to portray a scene typical in anime, rather than a landscape, that’s clearly transformative and derivative, but not substantially similar. If you use the style to make prints of waves breaking around Mt. Fuji, that’s substantially similar. So like, as to dude’s anime style, if you use the same style to make landscapes, certainly that’s not infringing, as it’s not substantially similar.
I also don’t see the threatening outcome the author suggests as worrisome. There are still exceptions for blatant copying that apply, mainly parody and fair use.
Will you guys shut up about this?
There are genuinely some big issues with AI that need to be addressed but they are drowned out by morons melting down over people making dumb little Ghibli style images for their own amusement.
Shout about insurance companies using AI to auto dent people’s medical claims, not about some dude Turnjng a picture of his cat into anime style
Its attacking on a cultural front and we will move on in a week. People still care more about insurance companies, trust me.
There is another aspect of this also. I could generate Ghibli style images a few years ago using better image generation models like stable diffusion or Midjourney. OpenAI is so lagging behind in terms of image generation it is comical at this point. But they get all the media coverage for these things as if they are inventing something out of thin air.
Most governments ignored the IP issues when other models were already doing these violations. Professionals are not using OpenAI. OpenAI only makes it so that these products reach big audiences. Then they become extremely accessible with the downside being that they are dumbed down. Thus, losing a lot of functionality.
OpenAI is so lagging behind in terms of image generation it is comical at this point.
They dropped a new image model last week using 4o to contextualize the request, it’s very very good. However it’s for paid subscribers only right now I believe.
However as you mentioned Stable diffusion and mid journey probably still have more customizability.
OpenAI is so lagging behind in terms of image generation it is comical at this point.
You’re the one lagging behind. OpenAI’s new image model is on a different level, way ahead of the competition
How so?
- Autoregressive model
- Multimodal with the LLM
- Can keep consistency between images
- Much better at text rendering
- Can combine images, like you have one image and you upload a picture of a jacket and say “put this on him” and it does it
- Can upload a picture of yourself and say “put me on the beach”, and then for example if you don’t like it you can tell it to do a different type of beach, and then say “and put me on a white horse and give me some nice beach wear” for example.
It understands what you’re telling it, and can generate images from vague descriptions, combine things from different images just by telling it, modify it and understand the context - like knowing that “me” is the person in the image, for example.
Edit: From OpenAI - “4o image generation is an autoregressive model natively embedded within ChatGPT”
Okay so how does that compare to whatever competition you’re referencing
No other model on market can do anything like that. The closest is diffusion based where you could train a lora with a person’s look or a specific clothing, then generate multiple times and / or use controlnet to sorta control the output. That’s fast hours or days of work, plus it’s quite technical to set it up and use.
OpenAI’s new model is a paradigm shift in both what the model can do and how you use it, and can easily and effortlessly produce things that was extremely difficult or impossible without complicated procedures and post processing in Photoshop.
Edit Some examples. Try to make any of this in any of the existing image generators
- https://www.reddit.com/r/ChatGPT/comments/1jl36h6/gpt_was_also_able_to_help_me_make_a_comic_ive/
- https://www.reddit.com/r/ChatGPT/comments/1jkl5m2/i_work_in_ecommerce_the_new_gpt_image_update_has/
- https://www.reddit.com/r/ChatGPT/comments/1jlewya/by_god_what_have_i_done/
- https://www.reddit.com/r/ChatGPT/comments/1jm8ddg/im_not_the_first_to_figure_this_trick_out_am_i/
- https://www.reddit.com/r/ChatGPT/comments/1jjsfkb/starting_today_gpt4o_is_going_to_be_incredibly/
- https://www.reddit.com/r/ChatGPT/comments/1jn2kpy/i_created_a_character_with_chatgpt_and_send_her/
- https://www.reddit.com/r/ChatGPT/comments/1jkaaxh/gpt4o_image_generation_is_absolutely_insane/
All diffusion and language models are autoregressive. That just means that the output is fed back in as input until the task is complete.
With diffusion models this means that it is fed an image that is 100% noise and it removes some small percentage of the noise and then then the denoised image is fed back in and another small percentage is removed. This is repeated until a defined stopping points (usually a set number of passes).
Combining images and using one image to control the generation of another has been available for quite a while. Controlnet and IPAdapters let you do exactly that: ‘Put this coat on this person’ or ‘Take this picture and do it in this style’. Here’s an 11 month old YouTube video explaining how to do this using open source models and software: https://www.youtube.com/watch?v=gmwZGC8UVHE
It’s nice for non-technical people that OpenAI will sell you a subscription in order to access an agent that can perform these kinds of image generation abilities, but it’s not doing anything new in terms of image generation.
I know them, and used them a bit. I even mentioned them in an earlier comment. The capabilities of OpenAI’s new model is on a different level in my experience.
https://www.reddit.com/r/StableDiffusion/comments/1jlj8me/4o_vs_flux/ - read the comments there. That’s a community dedicated to running local diffusion models. They’re familiar with all the tricks. They’re pretty damn impressed too.
I can’t help but feel that people here either haven’t tried the new openai image model, or have never actually used any of the existing ai image generators before.
you know enough about the model for me to immediately distrust your opinion on the matter. why don’t you head back to ycombinator or whatever hole you crawled out of
OpenAI picked Studio Ghibli because Miyazaki hates their approach.
I highly doubt it. They picked it because the Ghibli style is very popular among users. There’s also no reason to believe that it violates “democratic values”. Since it’s popular, the general population is voting that they LIKE it, not that they oppose it.
Downvote me all you like, but this is trying to put a lot of malice where the simpler explanation is just “money”.
Yeah it’s not like this is the only way to generate the style, it’s relatively simple to even do it locally. It’s just popular
Yeah the text makes many freestyle assumptions, although the overall sentiment is correct that these big companies and especially egocentric billionaires do stuff to trigger others simply for power display. I believe the text linked about it being a distraction for the new round of funding is the real reason.
What kind of article is this? They misattributed a quote, then admitted the misattributed the quote, then doubled down on it, and then threw in a political message.
People, this is rage bait. It’s yellow journalism. Don’t fall for this shit.
What quote is misattributed? Also it appears to be a blog post, I don’t really think its intention is to report on the facts but rather provide analysis. Fuck OpenAI for this and many other things, the ire is well deserved.
They give the Miyazaki quote and then say, “of course, he wasn’t talking about generative AI, but he could have been.”
That’s not what misattributed means especially regarding a quote. It would be misattributed if they said someone else’s name. Anyways how is it wrong (or whatever you meant) to say that what he’s saying about an older version of similar tech is applicable to a newer iteration? Either way this isn’t a news article, it’s a blog post. Who cares if it’s editorialized?
Either way this isn’t a news article, it’s a blog post. Who cares if it’s editorialized?
People who would rather hear the truth and not fancy lies that appeal to the masses.
Okay. Have you tried looking elsewhere than a blog post that never claimed to be “the truth”?
Anyways that’s a garbage argument. I’d like to know how you’ve been managing not to find anything opinion based in whatever corner of the internet you’ve come from. If you’re only willing to see things that are anywhere near “the truth” you should be reading an academic publication, not social media.
I don’t get my news from tante.cc
But the fact that I don’t use them for my news doesn’t mean that they’re not lying (“editorializing”) for profit, which is a bad thing for everyone who cares about not being misinformed since people, who do read trash like this, use this kind of ‘news’ as the basis of their opinions.
That linked X post from the White House at the end leaves me speechless.
Utterly inhumaneWe as the people of the united States have to do something. If you aren’t part of a movement yet join one, anyone, most of them are communicating with each other at this point.
Is this fashion comeback ? Style transfer was popular 10 years ago.
I feel like they’re reading too much into this.
Nah information should be free. Ghibli doesn’t own its style. Fuck this copyright propaganda machine.
Sucks because ghibli has always been really protective of its ip and in the future it maybe made harder and harder to watch it.
Ugh, why are they quoting that blowhard David Gerard
Edit: Lotta David Gerard fans here
https://www.tracingwoodgrains.com/p/reliable-sources-how-wikipedia-admin
I see it as enabling people to make images in a style they admire and would like to draw but don’t personally have the skill. To me the concept of copyright is the only difference between AI art generators and say, springy leg braces that let you slam dunk like Kareem Abdul Jabbar. I understand there are business ramifications some people might object to, but I don’t get the moralistic part of the outrage. Maybe somebody can help me understand by explaining it rationally without screaming or calling me names, but spitting rage at me is pointless.
edit: from the abundance of downvotes and lack of explanation I take it people know they’re supposed to be outraged but don’t know why. The telltale mark of meme culture, wear it proudly!
The moralistic outrage is that people still have an outdated concept of intellectual property, and a blanket fear of corporations owning technological progress.
The truth is, no one can actually own an idea or style. But we have laws that try to make it a real thing. Because of regulatory capture, copyright truly only benefits corporations with lots of money, not all the little indie artists that actually would need it.
Hell, most these indie artists make their money drawing and selling fanart, which is the most literal definition of copying. Yet no one worries about that.
AI does not know or create anything. Without stolen training data what would your fancy LLM actually be able to do?
It’s not my LLM, but like most software developers I admit I “stole” the same training data to learn programming.
Is it really a ‘move to allow’ style prompts? They’re just no longer preventing people from doing that.
It’s weird that people who profess to be staunch defenders of art don’t understand that stealing styles is fundamental to art. If enough people steal a specific style then art history just labels it a ‘movement’. Look on this page: https://magazine.artland.com/art-movements-and-styles/ and you can see that the thing they’re describing is a lot of people copying the same style.
Drum and Bass, a music genre, was essentially built on a “”“”“stolen”“”" clip from The Winstons in a song called Amen, Brother. The Amen break (you’ve certainly heard it even if you don’t know the name) is copied over and over and over.
This is just the latest social media trend trying to shoehorn issues into the ‘AI-bad’ meme. Stealing styles is not unusual or even immoral. It is literally the foundation of art.
This is just outrage farming, because 1. People are familiar with this style and 2. The primary artist who made the style popular is against AI.
