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

451–460 of 488 posts

Re: Stable Diffusion is a big deal

#451

Post by fleddr »

mattmanser wrote: post #247
nvln wrote: post #211

My son is an artist and we have conversations around AI generated art almost every week these days. Our general consensus has been: 1. Artists that use SD etc., will save a lot of time and have a big edge. Example: Generating references instead of searching for them, Generate random inspiration(s) for abstract ideas, Variations of a particular idea to compare and contrast. 2. Non artists who have powerful imagination…

So I just tried beta.dreamstudio.ai, and am a bit underwhelmed.. It's compositions generally look silly or completely ignore the instructions, even if I turn cfg way up. I was generally generating 9 compositions at a time, here's some of the things I've tried: "An elf fighting an orc in a forest" Half the images had christmas elves so it's obviously bad at context, half looked like miniatures (models like games works…

Your experience is quite common, especially when new to prompting. You can get more control by checking prompt guides, but this still leaves a lot to be desired.

You seem to have missed the main point of the article this thread is based-on though. With img2img, you can input a kid-level MS paint input and control composition in detail. With inpainting you fix individual parts, with outpainting the surroundings/background. With text inversion, you can reuse successful parts in any context. People are even making the first videos and animations.

All of this is available today, not years from now. The workflows are still clumsy, but it's been a week. Many people are working on the UI to streamline all of this. You're severely underestimating progress.

Re: Stable Diffusion is a big deal

#452

Post by fleddr »

quitit wrote: post #246

I've said this ad nauseam - but people who think this is going to kill an industry clearly have no idea of said industry. It's a fantastically great tool, and a very exciting space, but reducing the function of creatives to people who draw pretty pictures is staggeringly ignorant.

You pose it like an all or nothing equation, which it isn't.

The way I see it, if you'd consider the art world a pyramid, the bottom is about to fall out. A lot of commercial artwork serves no deeper meaning but pretty decoration.

The emphasis will move to ideas instead of just execution. Artists will soon find out about the avalanche of people that have creative ideas yet can't draw or paint. They'll be unlocked.

Re: Stable Diffusion is a big deal

#453

Post by fleddr »

kjkjadksj wrote: post #355
zulban wrote: post #280

Earlier quoted context omitted.

Live music didn't disappear but the industry took a huge permanent culling when any cheap venue could just play a track. The replacement doesn't have to be great to trash an industry. It just has to be good enough for most. This is.

How about the fact that you can go on github today and find a thousand ml based music generators probably, and yet people still like going to concerts and seeing an artist play an instrument.

Not a meaningful comparison. Seeing an artist play an instrument is a fundamentally different experience. It's visual, much more impressive sound, a social experience, and so on.

With imagine generation you just look at output. There's no difference between seeing the output of a human-created image or an AI-created image, people can't tell.

Re: Stable Diffusion is a big deal

#454

Post by fleddr »

habosa wrote: post #121

These new models completely changed my mind about how much impact AI will have in my lifetime. They are the most impressive software achievements in decades and anyone who has a “meh” reaction will absolutely end up looking silly. I’m not optimistic that their impact will be positive though.

I will up you by saying that even your prediction is silly, for using the word "lifetime".

Not lifetime. Right now. Just one week of StableDiffusion in the wild has seen the development of many clients, GUIs, optimizations, plugins for commercial software, editing workflows, the list is endless.

The speed of it all is dazzling. It will not take long before this is assembled into a smooth experience that runs everywhere, with obvious end state being your phone.

And that's still just artistic image generation. The open sourcing of it means people can make any vertical, precisely optimized for a particular (commercial) domain.

Almost all shortcomings seem to be crushed in no time. Bad at drawing eyes? Here's a new encoder that fixes it.

The creator of StableDiffusion indicated to soon tackle music generation and if I remember correctly, even poetry. And there's 3D and video.

The ultimate end state: if you can imagine it, it can generate it. Not only are we much closer to that point than society realizes, we're also moving at an exponential speed towards it.

Re: Stable Diffusion is a big deal

#455

Post by fleddr »

andybak wrote: post #221
simultsop wrote: post #150

Are there any standards/frameworks emerging on how to write texts/prompts?

Not sure what you mean by "standards/frameworks". It's an art rather than a science. Some of the tips passed around smack somewhat of cargo-culting, some genuinely make a difference. Everyone has their own approach and some techniques work for one topic but not others.

There's even a brand new word for it: prompt superstition.

"8k, highly detailed, ultra super very detailed, epic lighting, octane render"

Re: Stable Diffusion is a big deal

#456

Post by andybak »

dannyw wrote: post #256
hackerlight wrote: post #235

Earlier quoted context omitted.

I misspoke. I meant to say photorealistic faces. I haven't seen a good one yet from DALL-E 2 or Stable Diffusion. The systems that are custom built just for faces do a good job at photorealistic faces, though.

They are bad because they explicitly filtered out faces from the training data.

Not true of SD

Re: Stable Diffusion is a big deal

#457

Post by throwawaylinux »

fingerlocks wrote: post #369
chucksmash wrote: post #364

Earlier quoted context omitted.

You're nitpicking one of the given examples without engaging the user on the point they were trying to make. To be nuanced, maybe they might have said, "cobblers are less in demand now that many people have moved from owning fewer pairs of shoes they make last through repair to owning more pairs of shoes that they tend to get rid of when they are worn out due to changes in construction materials used in production,"…

Yes, I nitpicked because OP was speaking in absolutes. The larger point that “technology destroys jobs” is not novel or interesting. But to know that these jobs still exist, and might be more common than you realize, now that’s interesting!

> Yes, I nitpicked because OP was speaking in absolutes. The larger point that “technology destroys jobs” is not novel or interesting.

That wasn't the larger point. And if that was your issue, why didn't you ask about it so I could have explained it to you better?

The point was that art as a profession and a commercial venture is not special or different from any other industry with respect to automation.

Re: Stable Diffusion is a big deal

#458

Post by thewanderer1983 »

noisymemories wrote: post #171
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.

This is patently, obviously _wrong_ for anyone who has tried learning any artistic skill in their life. Sorry to be this straightforward, but it gets on my nerves every time I read it. If you tried learning, let's say, the chiaroscuro technique from Caravaggio you'd be analyzing the way the painter simulated volumetric space by using white and dark tones in place of natural lighting and shadows. You wouldn't even thi…

>This is patently, obviously _wrong_ for anyone who has tried learning any artistic skill in their life. Sorry to be this straightforward, but it gets on my nerves every time I read it.

Both my parents (though retired now) were commercial artists. I was trying to be an artist at one point in my life before moving in Engineering and Science. All my parents friends are artists so I grew up around artists.

Ask any artist here who is using illustrator, Photoshop, Krita etc. How often do they google image search for textures, or reference images that gets incorporated into their artwork? The final artwork is their own but it may incorporate many elements from others artwork.

>If you tried learning, let's say, the chiaroscuro technique from Caravaggio.. You wouldn't even think of splitting the whole painting into puzzle size pieces while checking how many how those look similar when put close one another.

Ever seen hyperealistic pointillism?

Who are you to be the arbitrator of how an artists creates their work? Have you ever gone to a modern art gallery and seen all the different methods people use to create artwork?

Art is boundless and unique to each who creates it.

If a Artist uses a tool to create art, everyone agrees that is art. It could be a paint brush, clay, software on a computer etc etc. If an artist uses AI as a tool to create art then suddenly it's not art.

Re: Stable Diffusion is a big deal

#459

Post by astrange »

senectus1 wrote: post #95
astrange wrote: post #76

Earlier quoted context omitted.

If by emotion you mean faces then that’s just a lot of data to fit in the model. The rest is up to the skill of the pilot. You can do a lot with color palettes for instance.

no its more than that... there is no "atmosphere" to them. they're just clean images. Its hard to explain.

They are very clean looking, it could be an effect of SD's compression to be so small.

MidJourney's production model is good about dirtying things up, you could use it as a post processor. I think its default style of turning everything into fine art-cyberpunk-movie matte painting gets old though.

I've gotten some good atmospheric photorealism results out of DALLE, but currently trying to get outpainting to work to extend them and it's tricky.

Re: Stable Diffusion is a big deal

#460

Post by Siira »

peoplefromibiza wrote: post #326
hbn wrote: post #311

Earlier quoted context omitted.

I don't think the word "stolen" applies here, at least not any more-so than I stole from him when I looked at that image. I'm not sure about that website rehosting his image, but in terms of the model people are running on their computers, it's not like it literally includes the image that can be pulled back out of it. A computer looked at that image, and tweaked a very small amount of numbers within ~5GB of numbers…

> not any more-so than I stole from him when I looked at that image I spoke with the author, the author disagrees with you. The reason he consider that stealing is because they took that image from the online shop where the disclaimer specifically states "2022 © Magnetic Press LLC by Arenathemes. All Rights Reserved" without even trying to contact the copyright holder. https://store.magnetic-press.com/products/viewpo…

> It doesn't give me the right to republish a similar book putting together what I saw

And how useful has this policy been for the public?

Post reply on HN