How do you square this with literally the first image in the OP showing side by side GPT reproing copyrighted work? imo a good modern art project would be someone making a website that “archives” NYT articles by laundering them through GPT rather than using the archive link that everyone posts to get around the paywall. Even HN guidelines bend over backwards to allow bypassing the paywall by allowing these links.
Please show me a prompt that reproduces it. Also to pass this test, it has to be just as easy as right clicking "download image"
The images in the article are done in reverse. They find a prompt that shows a copyrighted character and then search for the matching image. That's not how piracy is done.
The author, I believe, is being purposefully deceptive and hoping people who don't use DALL-E see "animated sponge" generating a SpongeBob look-alike and think they should be burned.
Not what I was arguing and you’re not going to win many arguments with anyone who is paying attending by coming out of left field with only tangentially related demands.
This is really about replacement. The copyright holders in the content industry aren't really afraid of LLMs infringing on past copyright, but are terrified of it replacing them on future work, and there is absolutely no legal protection from this. The lawsuit might officially be about copyright, but that's just because it is their only available legal angle of attack.