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

Stable Diffusion is a big deal

simonwillison.net

Post reply on HN
Print view View original

61–70 of 488 posts

Re: Stable Diffusion is a big deal

#61

Post by astrange »

choeger wrote: post #56

> No-one expected creative AIs to come for the artist jobs first, but here we are! Maybe that's because we never really thought about it. In hindsight, it's only logical. For an artistic rendering, correctness doesn't matter much nor does understanding the model. For flying a plane or driving a car or just transforming code from one language to another, it very much does.

There are some kinds of “correctness” that matter. When you’re doing something new, you can only really do one new weird thing at once. Too many weird things and it falls apart or distracts the viewers.

(The “cow tools” Far Side comic is a dumb example.)

Since these AIs can’t count, that extra weird thing is probably going to be people with too many fingers. I actually have a personal collection of weird Midjourney images because if you ask for wide aspect ratios it starts generating multiples of the same thing and fusing them together…

Re: Stable Diffusion is a big deal

#62

Post by amval »

ryandv wrote: post #22

Quoted post unavailable.

Probably because: 1. You don't make any actual point that could lead to a constructive discussion. 2. What you are saying is not true: this model was developed in an university in Germany and sponsored by a private company. The only public founder I could find is this one: https://www.crunchbase.com/person/emad-mostaque

Perhaps you should reevaluate your own biases as well.

Re: Stable Diffusion is a big deal

#64

Post by dsign »

Fordec wrote: post #11
michaelchisari wrote: post #6

After using SD heavily for a week, I half agree with this. It is incredibly disruptive, and it's wild how much it accelerates the creative process. I'll give you that. But two things I've noticed: First, artists will still have a massive advantage over non-artists with this tool. A photographer who intimately knows the different lenses and cameras and industry terms will get to a representation of their idea much fas…

For every tool, a better output will always be created by people who specialize in a craft. Just as photoshop revolutionized photography, to this day you can tell the difference easily between and bad 'shops. In video games upon release everyone is bad and its an even playing field. But as people practice and gain experience they improve their usage, refine their approaches. You eventually see metas develop and best…

I agree, this would be an incredible tool. I can see how some of the outputs may help me improve a piece I'm working on, even if I would never use the model's output for my final product.

Re: Stable Diffusion is a big deal

#65

Post by astrange »

LewisVerstappen wrote: post #3
thot_experiment wrote: post #2

Maybe this current explosion in the relevance and visibility of this kind of AI model will finally lead us to rethink how insanely nonsensical our IP systems are. I'm not holding my breath, but there's hope that this sort of thing will (combined with situations like the HBO debacle) clarify the need for massive IP reform in the cultural zeitgeist. The problem here isn't that the model was trained on copyrighted works…

Why do people keep bringing up copyright anyway? It seems pretty clear that the images being generated by StableDiffusion are transformative so it's protected under Fair Use. Am I missing something here?

Online artists esp. fanartists have a strict moral system with rules like “credit the original artist” that isn’t based on actual laws, so they’re upset about this.

Re: Stable Diffusion is a big deal

#66

Post by abledon »

senectus1 wrote: post #37

I'm no artist.. though I'll admit i dabble, but is it just me or does everything that is generated in this seem to lack emotion ? I have seen some very impressive pictures but nothing that seems to "emote"

yeah one thing I think of is when an artist 'creates' something, its coming from their brain, like expressing what they feel. When AI generates the Art, even though there is a human curator, the lines and shapes the AI picks out is all random/arbitrary/no-emotion (even though it draws upon a huge dataset of lines/strokes/styles to form). It's "noise" art. The noise is just cleaned up to resemble real things.

Re: Stable Diffusion is a big deal

#67

Post by abhisuri97 »

I’m more trying to see what the utility of stable diffusion (or just the text to image problem) in the long term. Right now people can play around with making weird art pieces and maybe it will be integrated into design tools...but then what?

Eg with other AI problems out there I can see a potential application to medicine, self driving cars etc, but I just don’t see what the bigger goal of this is going to be.

Re: Stable Diffusion is a big deal

#68

Post by astrange »

trention wrote: post #59
thewanderer1983 wrote: post #55

Earlier quoted context omitted.

Artist today use the exact same method of learning from other peoples artwork to generate new artwork and styles. These models are learning just like any artist learns and then producing new content.

Even if that's true (which it isn't), it doesn't matter. It's absolutely logically consistent to allow humans to do it while forbidding AI to do it.

The problem is forbidding it won’t do much.

- It appears people can train AIs from scratch or at least fine-tune them at home.

- Even if your art isn’t in “the training set”, that does not prevent the AI from learning its style. (Someone can decode it to CLIP embeddings. It could have a really good text model trained on vivid art museum descriptions of your art.)

- The ability of an image model to generate your art means it could also be trained in reverse to recognize it, producing a caption model, which would give vision to the blind. And surely you’d feel bad about that.

Re: Stable Diffusion is a big deal

#70

Post by t-3 »

throwaway1851 wrote: post #21
klipt wrote: post #12

Earlier quoted context omitted.

> “Defendant built an algorithm that memorized features of Plaintiff’s IP. Defendant’s algorithm recombines parts of those features in order to produce works in the same domain that compete with Plaintiff’s work, all without Plaintiff’s consent.” The fun part is, this is how human artists learn too.

They absolutely do. What they don’t do is mechanistically clone compressed mathematical representations of input data. The human part of the creative process could very well be a distinguishing feature, legally.

Tracing and collage both mechanistically clone elements of the input. Trendfollowing, mimicry, and copying others is standard in any creative area.

I don't think computer-generated works will be easily distinguishable from human unless they're desired to be or shipped with metadata. It's already hard enough to distinguish human artists from other human artists without having names attached up front.

Post reply on HN