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

Stable Diffusion Public Release

stability.ai

Post reply on HN
Print view View original

171–180 of 437 posts

Re: Stable Diffusion Public Release

#171

Post by derac »

daenz wrote: post #30
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…

>This is one of the most important moments in all of art history. I agree, but not for the reasons you imply. It will force real artists to differentiate themselves from AI, since the line is now sufficiently blurred. It's probably the death of an era of digital art as we know it.

Making art is already not making much money for the vast majority of producers (outside of 3d modeling). There really aren't very many jobs in making art. I'd reckon most people are artists because they love doing it.

Re: Stable Diffusion Public Release

#172

Post by mattkevan »

andybak wrote: post #140
superdisk wrote: post #93

Just tried it, the results seem pretty poor, about on par with Craiyon/DALL-E Mini. I don't think OpenAI should be worried quite yet.

With all due respect, I've been using it for over a week and I don't think you've given it a fair shot. There's plenty of cases it's worse than Dall-E and there's plenty of cases where it's better. Overall it seems to show less semantic understanding but it handles many stylistic suggestions much better. It's definitely in the right ballpark. In fact I'm still using a wide range of models - many of which aren't regar…

Agreed. I still primarily use vqgan + clip, which is nowhere near state of the art, but produces really interesting results. I’ve spent a long time learning to get the best out of it, and while the results aren’t very coherent, it’s great at colour, texture, materials and lighting.

Re: Stable Diffusion Public Release

#173

Post by tgv »

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…

You’re right. They call this an “ethical release”, but what ethics, I may ask. A profitable IPO is more likely to have been their consideration. As other researchers before them, they are willingly releasing something with the potential to do harm, or pave the way for it, washing their hands in innocence.

Re: Stable Diffusion Public Release

#174

Post by gitfan86 »

fpgaminer wrote: post #111
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…

If I recall from the interview with Stability.ai's founder, he has more or less the same opinion, and that humans will adapt to the new technology as we always have. I figure "please don't abuse this technology" warning stickers are more CYA. It'll make the vast majority of judges look at a motion to dismiss and not blink an eye.

Historically, he is correct. It is easy to find people that were against TVs, Cars, Trains, Electric Cars. Those people were not entirely wrong with their logic. Trains and cars did make it much easier for scammers to come into a town and then leave quickly.

Re: Stable Diffusion Public Release

#175

Post by squeaky-clean »

badsectoracula wrote: post #44

Is there any way to download this on my PC and run it offline? Something like a command-line tool like $ ./something "cow flying in space" > cow-in-space.png that runs with local-only data (i.e. no internet access, no DRM, no weird API keys, etc like pretty much every AI-related application i've seen recently) would be neat.

There's some additional discussion on running it locally here

https://old.reddit.com/r/StableDiffusion/comments/wuyu2u/how...

Re: Stable Diffusion Public Release

#177

Post by Geee »

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 think the actual problem is that this gives plausible deniability against photographic evidence, which might result in increase of bad behavior. Even cameras which cryptographically sign their output can't prove that the input was actually photographed from the real world, or if it's just an image of an image.

Re: Stable Diffusion Public Release

#178

Post by derac »

jmfldn wrote: post #69
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?

This Twitter thread has a lot of great comparisons of the models with different prompts.

https://twitter.com/fabianstelzer/status/1561019215754280963

Re: Stable Diffusion Public Release

#179

Post by scarmig »

suby wrote: post #73
gjsman-1000 wrote: post #55

Earlier quoted context omitted.

I had a thought for a utopian/dystopian world. One day these image generators will also get video support... and pornography support. When that happens, a few things may occur that I think are reasonable to predict: EDIT: Original post was way too wordy, TL;DR: When AI-generated pornography becomes available, it could be likely that demand for "real" pornography disappears because the AI will match and surpass the "r…

There will be a huge and unkillable market for non-AI generated pornography, even if people cannot tell the difference in an AB test. The demand will be too strong and I don't think there will be much outcry to ban it if it's all consenting adults.

If people can't tell the difference in an AB test, how will the real porn out compete the generated stuff? Porn distributors aren't known for their truth in advertising or care in sourcing of material. And even if they were, how would the porn distributors be able to source the real stuff, when anyone can create porn of anyone they imagine? You might say PKA will save us, but people aren't going to be typing out `gpg --verify` when their hands are otherwise occupied.

Re: Stable Diffusion Public Release

#180

Post by Rzor »

barbecue_sauce wrote: post #169
criddell wrote: post #136

Earlier quoted context omitted.

Would a Mac Pro with a Radeon Pro W5700X with 16 GB of GDDR6 memory work?

It says that NVIDIA chips are recommended but that they are working on optimizations for AMD. This implies to me that it probably involves CUDA stuff and getting it to run on a Radeon would be potentially difficult (I am not an expert on the current state of CUDA to AMD compatibility, though).

AMD's answer to CUDA is called ROCm. I've been doing a little research on it since a few weeks ago and it seems to be funky when not outright broken. It's absolutely maddening that after all this time AMD doesn't have proper tooling on consumer GPUs.
Post reply on HN