Stable Diffusion Public Release • Hacker News Board

Re: Stable Diffusion Public Release

#61

Post by digitallyfree » Mon, Aug 22, 2022, 7:01 PM UTC

fluidcruft wrote: post #21

cercatrova wrote: post #6

I've been looking forward to this. The license however strikes me as too aspirational, and it may be hard to enforce legally: > You agree not to use the Model or Derivatives of the Model: > - In any way that violates any applicable national, federal, state, local or international law or regulation; > - For the purpose of exploiting, harming or attempting to exploit or harm minors in any way; > - To generate or dissem…

It seems like the sort of things you would require so that HuggingFace doesn't get roped in as defendants in lawsuits related to things that others do with the code. So if for example someone builds something that generates medical advice and gets sued for violating FDA requirements or damages or whatever then HuggingFace can say that was not something they allowed in the first place.

I'm actually seeing these types of conditions becoming more common in software EULAs as well, as an boilerplate addon to the usual copyright notices and legal disclaimers. Don't have examples off the top of my head but I've seen clauses that this application may not be used for the enablement of violence, for discriminatory purposes, and so forth. It really is a CYA sort of thing.

Re: Stable Diffusion Public Release

#62

Post by hedora » Mon, Aug 22, 2022, 7:02 PM UTC

cercatrova wrote: post #6

I've been looking forward to this. The license however strikes me as too aspirational, and it may be hard to enforce legally: > You agree not to use the Model or Derivatives of the Model: > - In any way that violates any applicable national, federal, state, local or international law or regulation; > - For the purpose of exploiting, harming or attempting to exploit or harm minors in any way; > - To generate or dissem…

Each of the “for any use ... which has the effect of ...” clauses probably bars any wide distribution of the output of this model.

Trivially: People have phobias of literally everything.

They ban using it to “exploit” minors, presumably that prevents any incorporation of it into any for-profit educational curriculum. After all, they do not define “exploit”, and profiting off of a group without direct consent seems like a reasonable interpretation.

I am not a lawyer, but I wouldn’t dream of using this for commercial use with a license like this. This definitely doesn’t meet the bar for incorporation into Free or Open Source Software.

Re: Stable Diffusion Public Release

#63

Post by ForrestN » Mon, Aug 22, 2022, 7:04 PM UTC

mabbo wrote: post #33

The most interesting part, to me, of a release like this is the amount of "please don't abuse this technology" pleading. No licence will ever stop people from doing things that the licence says they can't. There will always be someone who digs into the internals and makes a version that does not respect your hopes and dreams. It's going to be bad. As I see it, within a couple years this tech will be so widespread and…

I wonder/hope if a downstream effect of this technological change will be the end of the idea of a "shameful" or "humiliating" image. We all have bodies, we all have sex, and I agree that soon images of ourselves being nude/having sex etc. will proliferate because they'll be generated instantly via Siri shortcut as part of casual banter.

In a world where every celebrity is having sex with gorillas, doesn't such an image lose its charge? Will Norms and values around sex/body shaming change?

Re: Stable Diffusion Public Release

#64

Post by boredemployee » Mon, Aug 22, 2022, 7:04 PM UTC

egypturnash wrote: post #41

mkaic wrote: post #2

This is one of the most important moments in all of art history. Millions of people just got unconditional access to the state-of-the-art in AI text-to-image for absolutely free less the cost of hardware. I have an Nvidia GPU myself and am thrilled beyond belief with the possibilities that this opens up. Am planning on doing some deep dives into latent-space exploration algorithms and hypernetworks in the coming days…

It’s the opposite of exciting if art is your job, lemme tell you.

Yep. I'm glad that didn't hit other spaces, like music, _yet_

Re: Stable Diffusion Public Release

#65

Post by h2odragon » Mon, Aug 22, 2022, 7:04 PM UTC

bluejellybean wrote: post #39

While neat, and no doubt impressive, it still utterly fails on prompts that should be completely reasonable to any sane human being/artist. Take something like "A cat dancing atop a cow, with utters that are made out of ar-15s that shoot lazer-beam confetti". A vivid description should be aroused in your head, and no doubt, I could imagine an artist have a lot of fun creating such a description... Alas, what the mode…

ITYM "udders"

also try "teats"

Re: Stable Diffusion Public Release

#66

Post by fezfight » Mon, Aug 22, 2022, 7:04 PM UTC

egypturnash wrote: post #41

mkaic wrote: post #2

This is one of the most important moments in all of art history. Millions of people just got unconditional access to the state-of-the-art in AI text-to-image for absolutely free less the cost of hardware. I have an Nvidia GPU myself and am thrilled beyond belief with the possibilities that this opens up. Am planning on doing some deep dives into latent-space exploration algorithms and hypernetworks in the coming days…

It’s the opposite of exciting if art is your job, lemme tell you.

Kinda like the creation of copilot and it's ilk for programmers, and gpt3 for writers. Ive seen some talk recently around 'prompt engineers'... Probably to some extent, every job will become prompt engineering in some way.

Eventually I suppose the AIs will also do the prompts.

At which point I hope we've all agreed to a star trek utopia, or it's gonna get real bad. Or maybe it'll get way better.

Re: Stable Diffusion Public Release

#67

Post by api » Mon, Aug 22, 2022, 7:05 PM UTC

"This release is the culmination of many hours of collective effort to create a single file that compresses the visual information of humanity into a few gigabytes."

If something like this is possible, does this mean there's actually far less meaningful information out there than we think?

Could you in fact pack virtually all meaningful information ever gathered by humanity onto a 1TiB or smaller hard drive? Obviously this would be lossy, but how lossy?

Re: Stable Diffusion Public Release

#68

Post by rvz » Mon, Aug 22, 2022, 7:06 PM UTC

hunkins wrote: post #3

This release changes society forever. Free and open access to generate a hyper-realistic image via just a text prompt is more powerful than I think we can imagine currently. Art, media, politics, conspiracy theories; all of it changes with this.

Yes. This changes everything.

Until this point, you really cannot believe any image you see on the internet anymore.

Re: Stable Diffusion Public Release

#69

Post by jmfldn » Mon, Aug 22, 2022, 7:07 PM UTC

simonw wrote: post #4

If you want to try it out this seems to be the best tool for doing so: https://beta.dreamstudio.ai/

I just tried it via this link. I'm not sure what I'm looking at here but the results were extremely underwhelming. I've used Dall E 2 and Midjourney extensively so I know what they're capable of. Maybe I'm missing something?

Re: Stable Diffusion Public Release

#70

Post by eastbound » Mon, Aug 22, 2022, 7:08 PM UTC

croes wrote: post #59

skybrian wrote: post #47

Earlier quoted context omitted.

It seems inevitable like it's "inevitable" that they Photoshop your face onto porn. Yes, of course it will happen but maybe not to most people? I'd guess inevitable for many celebrities.

It will be more now. Photoshop needs some skills. With AI it gets easier and easier.

But even today, we deal correctly with it. Fakes and real photos are mingled together in 9Gag/LatestNews reports about Ukraine. Under the fakes (and the real), people ask for confirmation. Someone says it’s true, no-one believes him, until a link to a newspaper is dropped. And 9Gag isn’t the highest IQ community around, so yes, general population does distrusts photos by default until proven.

They are laughed at anyway if they tell a story coming from a forged photo.

Sure, newspapers could forge stories, display pictures with, I don’t know, Biden’s son with a crackpipe, and make the populace believe untrue stories. But guess what, they already do it anyway, newspapers already “spin” (as they say, i.e. forge, suggest without literally saying) stories all the time.

The world deals correctly with fakes.