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

Stable Attribution

stableattribution.com

Post reply on HN
Print view View original

271–280 of 365 posts

Re: Stable Attribution

#271

Post by pksebben »

dahart wrote: post #133
bioemerl wrote: post #94

Earlier quoted context omitted.

I don't get the assumption that there should be. The machine was trained on publicly available hard that was already free. Why do people think they need to be compensated for something they put up online for free? They don't have the right to not allow people to learn from it, that's just never been a part of copyright.

Available for free online is not a valid justification for copying under copyright law, though, right? You can’t distribute something just because you can see it. True for museums and magazines as it is for online content. > They don’t have the right to not allow people to learn from it, that’s just never been a part of copyright. Yeah this is true. Stable Diffusion and other neural networks are not “learning” from i…

it's not remembering pixels. for it to do that, it would have to have the pixels stored somewhere. It does not.

The laion 5b dataset is in the neighborhood of 220TB. (1) That is how much storage space you need to remember the pixels.

The stable diffusion 1.5 checkpoint is 7gb. (2)

1 https://github.com/rom1504/img2dataset/blob/main/dataset_exa...

2 https://huggingface.co/runwayml/stable-diffusion-v1-5/tree/m...

Re: Stable Attribution

#272

Post by HammadB »

kmeisthax wrote: post #144
minimaxir wrote: post #120

Earlier quoted context omitted.

> The AI does NOT build or use reference boards for specific prompts. The only reference it has is the prompt itself, which gets distilled down into a list of 512 numbers, each one of which the AI associates with a particular image feature (or set of features). Minor technical clarification: For SD 1.X, CLIPText encodes a prompt and passes a (77, 768) [edit: up to 77] matrix to the core UNet. For SD 2.X, OpenCLIP pas…

Huh. I thought those got projected into the (512) shared CLIP space before getting passed to the conditional blocks.

Nope, this is why you cannot use images as prompts without some workarounds! SD doesn’t use the shared CLIP space but the text encoded before projection

Re: Stable Attribution

#273

Post by pksebben »

bioemerl wrote: post #181
adriand wrote: post #160

Earlier quoted context omitted.

Artists shouldn’t have to compete against themselves . If these AI companies didn’t use the work of artists, the output would suck! I’m sure many artists would be happy to compete against the artistic talents of software developers. But they’re not, they’re having to compete against their own work. I listened to an interview with an artist who referenced the “three C’s”: Consent Credit Compensation These seem reasona…

The request is to eliminate the technology. The training data set for stable diffusion has 5 billion images. Even if it was a single dollar per image, the data set as a whole would cost $5 billion. Getting consent of all the authors within that 5 billion images, managing the infrastructure of paying them, would be a Herculean task, and the cost of that task would far outrun the 5 billion spend if you gave a dollar fo…

This. I see and understand the FUD on behalf of the artists. It's real; this is going to change things in a way that makes their lives harder, and that sucks.

What I see in the discussion is a results first, truth second thing. This is understandable - in an existential fight, damn the consequences - I'm swinging for my own survival.

What's missing from that analysis though, is that by constricting open source models, you're not by any measure stopping the development or deployment of these models. Google will make one. Microsoft will make one. Apple will make one.

Adobe will make one. Then, in order to use these technologies, you will have to pay. A lot - and it won't be paying the artists. Sure, the initial training night send out a few cents for each work included, a small price for Google to pay to prevent competition. But you're smoking the wrong shit if you think for a second that won't change as soon as the lead is cemented.

So now you're still looking for another job, and you don't even get to play with this new tech and continue making art, because you don't work for Google.

Re: Stable Attribution

#274

Post by sinuhe69 »

Haha, I just uploaded a photo of mine to test and SA promptly reported a dozens supposedly human-made "sources" for my photo! I find it hilarious! No, that's not how attribution should work.

Re: Stable Attribution

#275

Post by nextlevelwizard »

dns_snek wrote: post #262
nextlevelwizard wrote: post #247

Earlier quoted context omitted.

How is what any of these image generators are doing any different from myself when I (try to) make art? I draw on my experiences and senses and try to reproduce a picture and those experiences include natural things I've seen as well as art others have made. More so how are these image generators any different from text generators like ChatGTP? I feel like if first tool out from these AI gates was a bot that wrote go…

> More so how are these image generators any different from text generators like ChatGTP? I've spotted this pattern a couple of times and this sort of circular reasoning seems concerning. Whenever one of (Stable diffusion, Copilot, ChatGPT) comes up in a discussion, their legitimacy seems to be swiftly justified by existence of the other two, even though they're all uniquely problematic in how they wash away attribut…

The thing is that no one seems to have these concerns with ChatGPT. When DALL-E came out it was touted as end of artists since now everyone could "just make their own art", but most people just see ChatGPT as a toy.

This is not circular reasoning, more observation of what people value. Since everyone can "Google and gather information" ChatGPT isn't valued enough, but since most people don't know how to draw DALL-E and Stable Diffusion are seen as industry destroyers.

Re: Stable Attribution

#276

Post by zerocrates »

grumbel wrote: post #130
mrtksn wrote: post #101

Earlier quoted context omitted.

They put it online for free for human consumption. Because the tech is very new, old concepts and laws don't cover it and that's not the point. The assumption that there must be compensation comes from the capitalist society that we live in. Switching to Communism or something else can be a solution to not directly pay people who do works and still have them around.

The Google bot has been consuming their art for many years, even copying it verbatim into Google's site and yet they didn't complain. This seems much more about the fear of competitions, than about violation of copyright. And yes, that's scary, but unavoidable as tech progresses. I don't think anything good could come from trying to strickten copyright here. The AI is obviously not copying directly, at best it takes…

People and companies have complained an enormous amount about Google's usage of images, particularly their inclusion in Google's site, and legal action or the threat thereof has caused Google to change how Google Images works before.

Re: Stable Attribution

#277

Post by okamiueru »

So, this is just a reverse image search, or does do anything more clever, like finding stronger matches in the latent space? For example the "style" could match, but a composition is completely different, etc.

So, even if it fails at correctly attributing source data. I'm wondering if it doesn't also fail at the concept of attribution. So far it just shows you some pictures with no attribution, and saying that whoever made those, made that.

Am I missing something? Why doesn't it know who the "human made sources" were made by?

Re: Stable Attribution

#278

Post by sebzim4500 »

okamiueru wrote: post #277

So, this is just a reverse image search, or does do anything more clever, like finding stronger matches in the latent space? For example the "style" could match, but a composition is completely different, etc. So, even if it fails at correctly attributing source data. I'm wondering if it doesn't also fail at the concept of attribution. So far it just shows you some pictures with no attribution, and saying that whoeve…

From what I can tell, it's just using the latent space to find similar images. Which is interesting, and potentially useful, but the fact that they are claiming that this is about 'attribution' puts it into scam territory IMO.

Re: Stable Attribution

#279

Post by gokuldas011011 »

romwell wrote: post #183
chii wrote: post #180

Earlier quoted context omitted.

> You do not* have the right to start a podcast where you recite the entire book from memory. which is fine - this is public broadcasting of the works, which is part of the exclusive rights of the owner. However, this is not what the AI is doing.

>However, this is not what the AI is doing. Yes, this is what the people who use the AI are doing. As well as people who release the AI model or a product that uses it.

So, that falls on particular outputs. Copyright infringements at the level per creations!

Re: Stable Attribution

#280

Post by bgidley »

chii wrote: post #171
wpietri wrote: post #151

Earlier quoted context omitted.

> But the point of good law isn’t to protect an established business model. Really? The Constitution specifically includes a bit about: "to promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries". That was quickly followed by the first copyright act. As I understand it, the copyright portion of this was e…

> the exclusive right exactly. This usage of works for _training_ is not part of that exclusive right, as far as i can tell. Otherwise, it would be a copyright violation for a human to read and learn off an existing works.

The question is how did you do training. If in the process you 'copied' the image (e.g. from network to memory) you did require copyright.

Reading with a human is not copying, but reading by machine is - there are several cases where that has been enforced. This is covered by reproduction rights.

Post reply on HN