Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

An IP attorney’s reading of the Stable Diffusion class action lawsuit

katedowninglaw.com

Post reply on HN
Print view View original

51–60 of 337 posts

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#51

Post by krisoft »

seydor wrote: post #27
acomjean wrote: post #18

>"The output represents the model’s understanding of what is useful, aesthetic, pleasing, etc. and that, together with data filtering and cleaning that general image generating AI companies do,2 is what the companies consider most valuable, not the training data.3" This didn't make any sense to me. Without the curated training data (images) how are they making the models? No matter what, putting images into your mach…

They are never going to compensate the artists. It's cheaper to hire 1000 designers to make 100000 images of artistic styles they are going for

uhm. What you are proposing is compensating the artists. Those 1000 designers are the artist in question then.

> It's cheaper to hire 1000 designers to make 100000 images of artistic styles they are going for

I bet that you are massively underestimating the cost of that.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#52

Post by gugagore »

seydor wrote: post #27
acomjean wrote: post #18

>"The output represents the model’s understanding of what is useful, aesthetic, pleasing, etc. and that, together with data filtering and cleaning that general image generating AI companies do,2 is what the companies consider most valuable, not the training data.3" This didn't make any sense to me. Without the curated training data (images) how are they making the models? No matter what, putting images into your mach…

They are never going to compensate the artists. It's cheaper to hire 1000 designers to make 100000 images of artistic styles they are going for

I think you underestimate the scale of data these models are trained on by many orders of magnitude.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#53

Post by mk_stjames »

scotty79 wrote: post #2

> Stability AI has already announced that it is removing users’ ability to request images in a particular artist’s style and further, that future releases of Stable Diffusion will comply with any artist’s requests to remove their images from the training dataset. This is incredibly disheartening. Who knows how long will it take to progress the tech to the point where anyone will be able to train and run models unrest…

I think this is why it was very important that their first release of the model weights (the v1.4 model) was and will continue to be very important. While training an entire model from scratch requires an insane amount of data and compute right now (and will continue to be out of reach of individuals for some years I think), fine-tuning the model can be done on a consumer graphics card (they call it Dreambooth)... and I believe we will see further refinements that allow more and more useful tuning and features being built by individuals on top that original model, making it more and more powerful for specific uses. So even if future efforts change the functionality, or nerf the output in some way, there will always be people developing tools on the original weights. That cat is out of the bag.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#54

Post by antiterra »

scotty79 wrote: post #2

> Stability AI has already announced that it is removing users’ ability to request images in a particular artist’s style and further, that future releases of Stable Diffusion will comply with any artist’s requests to remove their images from the training dataset. This is incredibly disheartening. Who knows how long will it take to progress the tech to the point where anyone will be able to train and run models unrest…

ChatGPT has shown that attempting to train the model to decline certain types of requests is of limited effectiveness and readily circumvented. The ability to make custom checkpoints and tools like Dreambooth will further limit these restrictions.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#55

Post by haswell »

shanebellone wrote: post #43

I've been saying this since it came out... Stable Diffusion is equivalent to hip-hop sampling in the 80s and 90s. The outcome is obvious.

I’ve heard this argument on numerous occasions but I have never heard someone justify it or why they believe it.

Are there specific similarities that make you believe these are equivalent scenarios? Not just “it feels thematically similar”.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#56

Post by belter »

matheusmoreira wrote: post #45
healsdata wrote: post #5

Earlier quoted context omitted.

I'm not sure I understand the point you're making. Its disheartening that artists can opt-out of having a computer algorithm make derivative versions of their creations? I'm probably on the opposite side of the fence. I do find it disheartening that it's opt-out instead of opt-in. The training set should be limited to public domain and CC-0 until such a time it can comply with attribution; then other CC works could b…

Yes, it is disheartening. Technology shouldn't be held back by this copyright nonsense. Public domain? Come on. Public domain barely exists anymore with the modern multicentury copyrights. What we need is enough computation power to run these models on our own computers, on our phones even. Then we'll be able to do whatever we want and there's nothing they can do about it.

Are you trolling? Because I can't guess...

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#57

Post by antiterra »

aliqot wrote: post #13
scotty79 wrote: post #2

> Stability AI has already announced that it is removing users’ ability to request images in a particular artist’s style and further, that future releases of Stable Diffusion will comply with any artist’s requests to remove their images from the training dataset. This is incredibly disheartening. Who knows how long will it take to progress the tech to the point where anyone will be able to train and run models unrest…

Where are all these self trained artists who learned their craft in a bubble, devoid of outside influence from other artists? Is it because there's a better paintbrush now, or is it because that paintbrush is not the 'real' way? This reminds me of the backlash against the wacom community on deviantart in the early days.

A computer program is not a person, so the argument that stable diffusion does what a person does is of limited relevance.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#58

Post by dns_snek »

healsdata wrote: post #5
scotty79 wrote: post #2

> Stability AI has already announced that it is removing users’ ability to request images in a particular artist’s style and further, that future releases of Stable Diffusion will comply with any artist’s requests to remove their images from the training dataset. This is incredibly disheartening. Who knows how long will it take to progress the tech to the point where anyone will be able to train and run models unrest…

I'm not sure I understand the point you're making. Its disheartening that artists can opt-out of having a computer algorithm make derivative versions of their creations? I'm probably on the opposite side of the fence. I do find it disheartening that it's opt-out instead of opt-in. The training set should be limited to public domain and CC-0 until such a time it can comply with attribution; then other CC works could b…

> I do find it disheartening that it's opt-out instead of opt-in

This is the crux of the issue for me. It's a different set of rules for AI companies than everyone else. If I started selling pirated copies of Nintendo games they would send an army of lawyers after me and this "opt-out" reasoning would not be a valid defense in court. These AI companies are trying to get away with stealing art and other content with a simple "whoopsie, we promise we won't do it again" when people demand that their own rights be respected.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#59

Post by nyolfen »

troyvit wrote: post #44
Karunamon wrote: post #38

Earlier quoted context omitted.

I hope the exact opposite. AI, including AGI if we ever get there, cannot be allowed to be strangled in its crib by artificially limiting the information it can learn from in the name of IP maximalism. IP law already goes way too far, the line should be drawn here.

If you want new art you probably want some form of IP. What's the incentive for an artist if at the first whiff of success their output is overtaken and resold by technocrats with machines?

totally ludicrous and hard to believe you actually think this. please take a glance at all of human history.

Re: An IP attorney’s reading of the Stable Diffusion class action lawsuit

#60

Post by philipwhiuk »

matheusmoreira wrote: post #45
healsdata wrote: post #5

Earlier quoted context omitted.

I'm not sure I understand the point you're making. Its disheartening that artists can opt-out of having a computer algorithm make derivative versions of their creations? I'm probably on the opposite side of the fence. I do find it disheartening that it's opt-out instead of opt-in. The training set should be limited to public domain and CC-0 until such a time it can comply with attribution; then other CC works could b…

Yes, it is disheartening. Technology shouldn't be held back by this copyright nonsense. Public domain? Come on. Public domain barely exists anymore with the modern multicentury copyrights. What we need is enough computation power to run these models on our own computers, on our phones even. Then we'll be able to do whatever we want and there's nothing they can do about it.

> Technology shouldn't be held back by this copyright nonsense.

The technology isn't.

The content is.

Post reply on HN