It's been amusing to watch everyone get big mad at every passing AI style fad. Some people can't stand seeing others have fun when that fun doesn't involve paying others.
By participating in the AI system you are contributing to the progress of the business and though you are not paying currency, you are paying with behavior that they can capitalize upon.
It doesn't require it, but even if some people are doing it on the paid plan - ignoring for the moment that the argument is asinine because this is a fixed recurring payment, not pay-per-use - why do you care? Some people have fun and it involves paying other people; if you're neither, it's still none of your business.
Why so defensive? I never stated any such opinion on it. I just pointed out the fact that it's not some sort of anti-capitalist activity that doesn't involve monetary transactions - billionaires are actively getting richer out of this.
> I just pointed out the fact that it's not some sort of anti-capitalist activity that doesn't involve monetary transactions
Nobody wants or cares about it being anti-capitalistic. It's just free snacks or up-front pay for an all-you-can-eat buffet. Yes, somebody pays for this, but no, this is not in scope or relevant for people enjoying the meal.
> billionaires are actively getting richer out of this.
As they should. Yes, we know they are, and it's fine. That is just the fundamental of market economy: someone earns money by providing useful goods or services. It's not a problem in itself, and neither is the fact that there may be a billionaire involved.
Getting wound up about billionaires specifically is a shtick of people who want to sell your eyeballs to advertisers, and/or sell you their ideology. The former is a cancer on modern society. The latter is even worse.
(Yes, there is growing income inequality, unfair concentration of wealth and power, and such. But putting it all on "billionaires" - the modern witches and demons - is just a trick that conveniently redirects any blame from people, organizations and systems contributing most to the core problem.)
It is quite ironic that OpenAI loosens its rules, promoted Ghibli images, apparently in direct opposition to the viewpoints of Ghibli's founder, while also pursuing DeepSeek for using OpenAI data without permission.
FYI, some of my further elaboration on the topic:
"Everything that you loved for its uniqueness, craftsmanship, cultural significance, will be mass-produced until you despise seeing it."
There are various ways to try and make sense of what we're living through right now, and commentary such as yours (presumably informed by decades of seeing this industry evolve) will be valuable in the coming years, as we all try to find our bearings in this post-LLM, post-truth, post-literate world.
It’s kind of interesting things like this are happening while at the same time, some leaders, like Musk, are making an anti-globalization argument because of the way that it pollutes a places culture.
Different leaders believe different things, and both of these guys will say literally anything if they feel its in their interest. I'm sure Musk is pro-globalization if globalization = Grok being the world's preferred LLM.
Its interesting that they picked an animation studio whose creator is famously opposed to AI, and whose stories frequently involve battles between technology and humanity.
It will specifically stop you if you try and ask it for the style of Disney or marvel, while Sama specifically tweeted an example of Ghibli. There is very clearly a conscious choice around which companies IP they protect, and which they're happy to exploit (eg, not US companies).
Like the "Her" tweet, I suspect this is another moment of Sama being bloody thoughtless and may very well come and bit him in the ass when he inevitably tries to deny this in the future.
EDIT: Okay, really weirdly, I tried this when this first came out and definitely seemed to hit a content policy block. and now it works.
I just asked it "Make an image of a karate fighter in the style of a disney animation" and it did it in distinct Disney style. What it produced did not look like public domain steamboat willy or anything either, more 60s era Disney.
Where are you getting this? It may have been after they reduced refusals recently and it wouldn't do it before?
Disney is also the Ghibli distributor in the US aren't they?
Strangely, if you're going to mess with IP in any country, Japan is one of the ones that you should not. Japanese law is infamously brutal about copyright and has no concept of fair use, so you can get smacked very easily by it.
I guess they don't have assets in Japan to risk and therefore don't care.
> I guess they don't have assets in Japan to risk and therefore don't care.
That's the benefit of being a corporation that keeps its physical presence in its home country, especially in the internet age. Perhaps the US gov could sponsor Starlink as a way to break local ISP bans…
Incorrect. I just asked ChatGPT to show an image in the style of Disney/Pixar, and another in the style of a Marvel superhero. Both worked without complaint.
I would share the conversation but it's not generating a link ("Sharing conversations with user uploaded images is not yet supported.")
Was just pointing that out since OP was saying it would refuse on Disney styles, though Disney doesn't have the copyright to Ghibli they still have US interests in the IP. But what he was saying doesn't seem to be true, it will do disney style drawings like it will do Ghibli.
(edit: from another reply it sounds like Disney is no longer the Ghibli distributor though, I guess that's why they are on HBO Max and not Disney+
Styles aren't copyrighted anyway, Steven Spielberg has produced Disney-style animated movies outside of Disney, and Disney has taken styles from others)
To back up the poster, Open AI loosened their rules. Earlier this week I did “dark wing duck as a knight” and it generated half and then said it couldn’t due to copyright. The same prompt now returns the image. Assume good faith instead of flying off and accusing people of being blind acolytes based on your personal idea of who they are.
This is an interesting point because it shows top-level control of output - by allowing one type of infringement rather than all types of infringement, it’s a human based decision. It’s only a matter of time before AI crosses into trademark domain issues and that’s where the big money lawyers will likely step in. As in, it harms the mark and it can be financially damaging unless stopped - and these systems clearly can be stopped from certain outputs.
Don't immediately believe what you read on Hacker News, certainly not to the point where you'd jump to such a significant conclusion over it.
ChatGPT is simply inconsistent about enforcing rules. Sometimes you ask it to generate an image and it says no, and then you ask it again in a slightly different way and it happily does it for you. One time I asked it to generate an image for me and it told me "I can't generate an image of Elon Musk but I can generate an image of someone who looks like Elon Musk. Let me know if you'd like me to do that instead", and the generated image looked exactly like Elon Musk.
I feel like the disgust that Miyazaki felt seeing an AI generated movie is very different than people editing their profile pics with a Ghibli style-transfer. (For several reasons, such as: the profile picture was already original art, no one is replacing artists with profile photo filters, and people are celebrating the style rather than profiting off plagiarism)
