Re:Judge Dismisses Copyright Claims Against AI Image Generators
by anon291 on Tuesday November 07, 2023 @12:46PM
(#38180138)
Can you explain in more detail what you want? If you undertrain a diffusion model it will produce meaningless gibberish. Which is still 'creative', just not what you expected. Training for a diffusion model is not 'inserting' images into a dataset. Rather you should view it as a 'taste' alignment.
Re:Judge Dismisses Copyright Claims Against AI Image Generators
by wredue on Tuesday November 07, 2023 @02:02PM
(#38181250)
>“If you under train a model, its results being copy and paste are fundamentally more obvious so I refuse to do that!”
Exactly dude.
Re:Judge Dismisses Copyright Claims Against AI Image Generators
by anon291 on Thursday November 09, 2023 @12:58AM
(#38201680)
But it's not copy paste. At no point does it not 'create' something from thin air. With training it produces things along the lines of what you want. That's the difference. It doesn't copy. It just notices what you like, and generates that.
Train an AI just on a few techniques. Then a couple fundamental stock photos with nothing but the subject. Then have it generate things outside of the realm of the stock photos recombined in a general style. Let us see your completely unique and unpredictable results.