Live data from Hacker News

Stable Diffusion is a big deal

simonwillison.net

291–300 of 488 posts

Re: Stable Diffusion is a big deal

#291
post #11

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…

Yes, but the consequence for society is that professionalization is coming to every single field, and the blue-collar jobs are evaporating, and we had better figure out how to get all those people either white-collar jobs or some other form of income if we want to avoid significant social instability.

It used to be that you could find work as a political cartoonist and draw a picture of some satirizable politician each day - maybe Dr. Oz with a funny-looking vegetable platter, or Biden falling off a bike labeled "build back better," or something - and that would be a career. Now it seems you can just type "Dr. Oz holding a vegetable platter, political cartoon" into one of these AIs and the bulk of the work is automated. Sure, you could spend the rest of the day refining it, but nobody's really looking for perfection or transcendent skill in their political cartoons.

You can still find work making political videos of Dr. Oz and vegetables (e.g. https://twitter.com/JohnFetterman/status/1564432981841907713). Today's image generation AIs cannot do videos like that. Tomorrow's will, before we know it, and again nobody's looking for more than a baseline level of quality there either.

And even the local newspaper that might have been employing a political cartoonist is being swallowed by high-capital-ownership companies that can replace a lot of the writing with AIs. The expectations of quality are a bit higher - though the AIs have mostly gotten the hang of sports writing - but again people don't expect The Smalltown Gazette to have the standards of The New Yorker.

Sure, the highest-quality drawings, feature films, and longform journalism will get a lot better. And that's great! But most people don't work on such things. What are they going to do? If nothing else, how will they remain an audience with money to spend on the highest-quality works?

(I am not advocating for stopping this process, to be clear. Smashing the AIs isn't a coherent proposal, and smashing the physical machines of automation didn't actually work when the Luddites tried. I'm advocating for admitting that this process is happening and drying up employment, admitting that keeping as much of humanity as possible under a good standard of living is important for humanity, and figuring out what to do about it.)

Re: Stable Diffusion is a big deal

#292

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…

> A photographer who intimately knows the different lenses and cameras and industry terms will get to a representation of their idea much faster than someone without that experience.

Sure. Only now the gap has been shrunk by a factor of infinity, because the time to representation was infinite for someone without that experience before.

Re: Stable Diffusion is a big deal

#293

Earlier quoted context omitted.

There's one huge difference between Copilot and stuff like this. Art that's 98% correct is awesome. Code that's 98% correct is completely useless. I think Copilot is going to live off hype for a while then tank and be looked back on as a failed experiment. Whereas I think that this kind of AI will eventually get to a point where it's extremely useful and could change up certain industries (game assets, marketing mate…

>Code that's 98% correct is completely useless. Whoa! Really? Seems like modern software releases would be excited to achieve 98% with the incredible amount of bug fixes/patches released very quickly after the massive beta test known as release day.

You are looking at code like a production line. It's a semantic construction.

Code 0.1% wrong, sends you to the Sun instead of the Moon, debits your account instead of crediting...

Re: Stable Diffusion is a big deal

#294
post #43
post #11

Earlier quoted context omitted.

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…

Have we seen similar behavior in the usage of older GPT-3 based tools, for example in copywriting or using Copilot? How's the story there?

There are genre fiction authors using GPT-3 based tools to generate prose for their 27th "merfolk romance" or "firefighting LitRPG".

Re: Stable Diffusion is a big deal

#295
post #291
post #11

Earlier quoted context omitted.

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…

Yes, but the consequence for society is that professionalization is coming to every single field, and the blue-collar jobs are evaporating, and we had better figure out how to get all those people either white-collar jobs or some other form of income if we want to avoid significant social instability. It used to be that you could find work as a political cartoonist and draw a picture of some satirizable politician ea…

> blue-collar jobs are evaporating

I've been hearing this for awhile, yet ironically right now we have the biggest shortage of blue-collar workers we've had in a long time.

Re: Stable Diffusion is a big deal

#296

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…

Anyone else like me say "2D art? Boring. The real world is 3D"?

Re: Stable Diffusion is a big deal

#297

Earlier quoted context omitted.

There's one huge difference between Copilot and stuff like this. Art that's 98% correct is awesome. Code that's 98% correct is completely useless. I think Copilot is going to live off hype for a while then tank and be looked back on as a failed experiment. Whereas I think that this kind of AI will eventually get to a point where it's extremely useful and could change up certain industries (game assets, marketing mate…

>Code that's 98% correct is completely useless. Whoa! Really? Seems like modern software releases would be excited to achieve 98% with the incredible amount of bug fixes/patches released very quickly after the massive beta test known as release day.

Code has to be 100% correct or else it’s considered a bug (assuming it’s syntactically valid).

Code that is 98% correct is actually much worse than no code at all. That’s the kind of code that will introduce subtle, systemic faults, and years later result in catastrophic failure when the company realizes they’ve been calculating millions of payments without sales tax, or a clever hacker discovers they can get escalated permissions by passing a well-crafted request etc.

Re: Stable Diffusion is a big deal

#298

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…

> A photographer who intimately knows the different lenses and cameras and industry terms will get to a representation of their idea much faster than someone without that experience. Sure. Only now the gap has been shrunk by a factor of infinity, because the time to representation was infinite for someone without that experience before.

Photographers also spend a great deal of time shooting real people and events. An AI-generated wedding photo is not a thing.

Re: Stable Diffusion is a big deal

#299

Playing with the prompt-only demo at https://huggingface.co/spaces/stabilityai/stable-diffusion I got the impression that many apparently harmless requests corner the model into a very sparse set of examples exhibiting extreme bias and utter nonsense. For example (seed 0, other advanced options at default values): Blue hamsters filling donuts with nails mostly edible donuts and quasi-donuts; some blue, but no hamster…

I think many of your prompts have trouble getting the right representation because you're using a lot of abstract and implied language. For these prompts to work right, you need to be very specific and often add keywords as if you're specifying a mood board. You also need some overlap between what the model has likely been trained on and the expected output.

The system works much better in img2img mode. Slapping together a rough shape of what you want in GIMP/Photopea and then applying a prompt to blend everything together make the entire process a lot more reliable.

Re: Stable Diffusion is a big deal

#300

Earlier quoted context omitted.

This is such a strange position to hold. At some point in the future may be confronted with an image that is emotionally stimulating, and you will have no idea whether a human or AI created it. Are you suddenly going to dismiss and discard it just because you subsequently learn it was AI generated? You must see how silly that is. Your "logic" for making AI art illegal is basically "don't like it". Your personal and s…

My position is perfectly logically consistent. Art is distilled human experience and human emotions emerging in a particular context after a chain of events. None of this is true of AI "art". What is not logically consistent is to claim that a black box utilizing statistical relationships between pixels in a giant dataset is an "artist" and that its products create "value". The compiler is not a programmer, AI can ne…

The chain of events:

Step 1: Tweak settings and type text

Step 2: Look at the result

Step 3: If you like the result, go to step 4, otherwise go back to step 1.

Step 4: Save and share the result

Feels like art to me. Ultimately it's still a human using a tool to create art. For me AI art is just the name of the art style.

Post reply on HN