OpenAI's move to allow generating "Ghibly stlye" images isn't just a cute PR stunt. It is an expression of dominance and the will to reject and refuse democratic values. It is a display of power
-
I think whats really happening behind the scenes is that the model you’re talking to makes a function call to another model that generates the image.
I haven’t seen it either so if you want that and don’t want to code it might be best to stick with paid, but something like that could easily exist somewhere else.
I bet you're right, but the fact that I never see it is a feature worth paying for, especially for a smooth-brain like myself.
-
Yeah they want corporations to own styles so the rich can be more powerful, the rich push this sort of propaganda out endlessly
This is just like china, copying stuff, or rather called steeling.
the original companies need to build their brand and style for decades and spend 100s of millions to improving to perfection.
then we have AI just copying it in matter of minutes.and you think 1 person should be able to steel all this work and legacy from 1000s of employees because its "protecting the rich"?
-
They're trying to make some type of argument that a private studio should have exlusive 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.
-
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.
As you have described the situation my question is if it would be similar to copyright Donald duck, despite not having drawn all possible poses and situations?
-
This is just like china, copying stuff, or rather called steeling.
the original companies need to build their brand and style for decades and spend 100s of millions to improving to perfection.
then we have AI just copying it in matter of minutes.and you think 1 person should be able to steel all this work and legacy from 1000s of employees because its "protecting the rich"?
first let's get something out of the way
the actual way that copyright works is that a few giant megacorps buy up everything and they end up owning copyrights to the vast majority of recognizable content.
so for example in 2019 over half of the movies released in theaters was owned by Disney. The same company that unilaterally has the ability to change US federal law when convenient for them.
studio ghibli is no different- they're a subsidiary of Nippon Television which has a $2B+ annual revenue
so keep in mind when you advocate here for stronger copyright protections, you are essentially saying that the biggest companies in the world deserve more money.
2nd- the "style" is not copyrightable. anybody can mimic the style. and guess what? if I make a cartoon and I make it look like studio ghibli style.. people are still gonna recognize it as "studio ghibli" style. they are basically getting free marketing. they are not losing out here.
-
first let's get something out of the way
the actual way that copyright works is that a few giant megacorps buy up everything and they end up owning copyrights to the vast majority of recognizable content.
so for example in 2019 over half of the movies released in theaters was owned by Disney. The same company that unilaterally has the ability to change US federal law when convenient for them.
studio ghibli is no different- they're a subsidiary of Nippon Television which has a $2B+ annual revenue
so keep in mind when you advocate here for stronger copyright protections, you are essentially saying that the biggest companies in the world deserve more money.
2nd- the "style" is not copyrightable. anybody can mimic the style. and guess what? if I make a cartoon and I make it look like studio ghibli style.. people are still gonna recognize it as "studio ghibli" style. they are basically getting free marketing. they are not losing out here.
I’m not entirely sure what this style even is - wouldn’t this same argument apply to Apple’s “Memoji” that has been out a few years?
-
As you have described the situation my question is if it would be similar to copyright Donald duck, despite not having drawn all possible poses and situations?
That's already the case. There would be two copyrights for a cartoon for Donald duck, and possibly, in fact likely, many others.
A copyright is essentially a right of enforcement. You don't have to register anything or file anything in order to gain that right. It's a right to sue someone to enjoin further use and potentially to recoup money damages if you can prove loss.
The standard for whether something is copyrightable at the outset is whether it is the product of a modicum of creativity, and reduced to a tangible medium of expression.
So far one cartoon of Donald duck, each drawn frame of the show would have its own copyright. Also, the character would have a copyright. The dialogue of the script would have another copyright. And the test for whether a particular character is something that can be copyrighted is to ask whether the character is separable from the overall work and whether the character is "well delineated."
Donald duck is certainly the product of creativity, it is reduced to a tangible medium of expression when it is drawn on paper, and it is the main character of the show and has its own personality and behavior. So it is pretty clearly of deserving protection. Although at this point in time, I believe some of Disney's earliest characters are now in the public domain, Even Mickey mouse, which people like my IP professor in law school said was never going to happen. This is because I believe in 1984 there was a law called the copyright act of 1984 but was colloquial referred to as the Mickey mouse copyright act. It was championed by Sonny Bono, who I believe was friends with Walt Disney personally, and which many said had the sole purpose of extending Mickey mouse's copyright for another 25 years or whatever it was. My memory is a little fuzzy on this. My professor figured that Disney was such a powerful institution that anytime Mickey mouse was about to fall into the public domain, Congress would stop it.
A doctrine sort of related to your question is called scen a faire. It is a French phrase which I have no doubt spelled wrong because I am on mobile. It means that elements essential to a scene of the kind which would be common to all scenes of that type, are not copyrightable. So this would include some background characters such as those that, despite being drawn in a creative way, are more so the product of the scene itself rather than any creativity. For example, if there is a scene in a cartoon where the character gets onto a train and hands the ticket to a ticket taker, the ticker taker character is probably not copyrightable.
-
No it isn't at all. Image to image "AI" is totally different from "AI" that denies insurance claims.
Nah dude this sounds to me lile a "duvide and conquer" strategy, make us believe that the grievances of one group contend with the grievances of another, quite the scab move.
-
Nah dude this sounds to me lile a "duvide and conquer" strategy, make us believe that the grievances of one group contend with the grievances of another, quite the scab move.
You can believe both are bad and complain about both while acknowledging they're separate things.
-
You can believe both are bad and complain about both while acknowledging they're separate things.
You can also acknowledge that both need their own space; If AI used for other nefarious purposes is so important to you your time is better spent making threads about that than implying threads about other issues are less serious.
-
You can also acknowledge that both need their own space; If AI used for other nefarious purposes is so important to you your time is better spent making threads about that than implying threads about other issues are less serious.
implying threads about other issues are less serious.
I never did this.
-
implying threads about other issues are less serious.
I never did this.
then this whole conversation is a little meaningless.
-
first let's get something out of the way
the actual way that copyright works is that a few giant megacorps buy up everything and they end up owning copyrights to the vast majority of recognizable content.
so for example in 2019 over half of the movies released in theaters was owned by Disney. The same company that unilaterally has the ability to change US federal law when convenient for them.
studio ghibli is no different- they're a subsidiary of Nippon Television which has a $2B+ annual revenue
so keep in mind when you advocate here for stronger copyright protections, you are essentially saying that the biggest companies in the world deserve more money.
2nd- the "style" is not copyrightable. anybody can mimic the style. and guess what? if I make a cartoon and I make it look like studio ghibli style.. people are still gonna recognize it as "studio ghibli" style. they are basically getting free marketing. they are not losing out here.
-
Anyone who doubles down this hard to defend AI art theft machines fucking hates human artists, who are a branch of intellectuals. Nazis are known to openly hate, abuse the rights of and mistreat intellectuals. Fuck kava.
You may or may not be correct in hating me but do not let my comments bring down the good name of kava
As for "doubling down so hard" I'd flip the message and ask you why you are simping for mega corps?
An AI is not doing anything a human wouldn't do. You look at a bunch of content. You learn from it and incorporate it in new synthesis.
It's not fundamentally different. So unless you can make a meaningful statement (beyond mild personal attacks) that illustrates the difference between the two, you will convince no-one
-
Do you know what the word unfettered means?
To answer your question, it's more about arguing for basic freedoms consistently than about arguing for disrespect.
When approaching these ethical questions, I think it's best to focus on the individual & moral reciprocity: should someone be able to express themselves in a way that offends me?
As long as it obeys the harm principle, the answer is yes.
Accordingly, anyone should be free to express themselves with imagery in the style of Ghibli (using tools such as AI) even if it offends the studio's founder, since it results in no actual harm.Since morality should be based on universal principles that don't depend on contingent facts of an agent (such as their characteristics), I find it clarifies questions to approach technology with their non-technological equivalents.
Would it be wrong to train a person to learn Ghibli art style so they could produce similar works in that style on demand?
The harm of that is unclear, and I would think it's fine.I don't see a general duty for a free society to fulfill a wish unless it's more of a claim right than a wish.
In particular, criticism is a basic part of art: a duty not to criticize artists (who wish not to be criticized) would be unjust.
While an artist should get credit (and all due intellectual property rights) for their work, once it's out in the wild it takes on a life of its own: people are free to criticize it, parody it, & make fair use of it.
Some wishes don't need to be fulfilled. -
To answer your question, it's more about arguing for basic freedoms consistently than about arguing for disrespect.
When approaching these ethical questions, I think it's best to focus on the individual & moral reciprocity: should someone be able to express themselves in a way that offends me?
As long as it obeys the harm principle, the answer is yes.
Accordingly, anyone should be free to express themselves with imagery in the style of Ghibli (using tools such as AI) even if it offends the studio's founder, since it results in no actual harm.Since morality should be based on universal principles that don't depend on contingent facts of an agent (such as their characteristics), I find it clarifies questions to approach technology with their non-technological equivalents.
Would it be wrong to train a person to learn Ghibli art style so they could produce similar works in that style on demand?
The harm of that is unclear, and I would think it's fine.I don't see a general duty for a free society to fulfill a wish unless it's more of a claim right than a wish.
In particular, criticism is a basic part of art: a duty not to criticize artists (who wish not to be criticized) would be unjust.
While an artist should get credit (and all due intellectual property rights) for their work, once it's out in the wild it takes on a life of its own: people are free to criticize it, parody it, & make fair use of it.
Some wishes don't need to be fulfilled.You really just wrote 4 paragraphs to say "I wanna make other people upset for my own personal gain at little to no benefit to the world at large"
-
You really just wrote 4 paragraphs to say "I wanna make other people upset for my own personal gain at little to no benefit to the world at large"
Not at all: logical ethical principles (golden rule, harm principle, freedom of the individual) & basic individual liberties in a free society.
Such a society where people are free to express themselves without doing actual harm is a benefit to the world "at large".
The alternative would be bleak. -
Not at all: logical ethical principles (golden rule, harm principle, freedom of the individual) & basic individual liberties in a free society.
Such a society where people are free to express themselves without doing actual harm is a benefit to the world "at large".
The alternative would be bleak.Okay, fine. I'll do big words with you.
Let's talk about the difference between Ethics and Morals. Ethics are what you individually follow, it's your Ethos. Morals are what a society follows.
It may be morally correct to say that "people should be allowed to do as they please" while also ethically correct to say "you shouldn't disrespect people's wishes."
You're right, that it is moral for someone to imitate art. At least, in our culture. But that's because most people follow the ethos that they can do and take as they please. I personally hold myself to a higher standard than that, and I hope you can do the same some day.
-
FWIW speech to text works really well on Apple stuff.
I’m not exactly sure what info you’re looking but: my gaming PC is headless and sits in a closet. I run ollama on that and I connect to it using a client called “ChatBox”. It’s got a gtx 3060 which fits the whole model, so it’s reasonably fast. I’ve tried the 32b model and it does work but slowly.
Honestly, ollama was so easy to setup, if you have any experience with computers I recommend giving it a shot. (Could be a great excuse to get a new gpu
)
Yeah, I think the Apple speech to text is pretty decent, but I think on ChatGPT they use the whisper API to return the text and it just seems to be a lot more reliable, especially when it comes to understanding random words in context
How much VRAM do you have on the 3060 to be able to fit the whole thing on the GPU?
-
Yeah, I think the Apple speech to text is pretty decent, but I think on ChatGPT they use the whisper API to return the text and it just seems to be a lot more reliable, especially when it comes to understanding random words in context
How much VRAM do you have on the 3060 to be able to fit the whole thing on the GPU?