Wednesday, September 14, 2022

Stable Diffusion is here to give you art on demand and to make artists obsolete.

For the past week or so, I've been playing around with Stable Diffusion. If you want to play around with it, you can go HERE to check it out. When you click on that link, it will take you to a link where you can enter a string of text and then hit a button and wait for a robot (an artificial intelligence) to generate an image based on those words. I did one last week where I wrote, "Timothee Chalamet as Legolas," and it generates some really uncanny valley images. It's kind of addicting to come up with a sentence and then wait for the A.I. to interpret your words and generate a bespoke image just for you.

Here are some images that were generated by other people who have been playing around with the software. 

It's definitely got some uncanny valley to it. Another thing it can do is generate images from other images. This second set of pics came from Reddit user argaman123. They created this image:
and then they added the following prompt:

A distant futuristic city full of tall buildings inside a huge transparent glass dome, In the middle of a barren desert full of large dunes, Sun rays, Artstation, Dark sky full of stars with a shiny sun, Massive scale, Fog, Highly detailed, Cinematic, Colorful

And then Stable Diffusion produced the following two images:

I haven't gotten anything close to results like this, but I'm still learning how the software works. It's also improving all of the time. I think it's only a month old at this point. However, anyone that does art for a living should be scared. Stable Diffusion literally makes images in less than five minutes. And as the software gets better, artists are probably going to become obsolete if this is the kind of quality we can expect from artificial intelligence.

The robots are coming for us all.


  1. I might have to check it out. Maybe I could use it for some covers. Maybe not people but backgrounds or something.

  2. Oh, that's what the brouhaha has been all about. I've seen headlines and tweets, but I wasn't sure what they were referring to.

  3. Playing with it for a couple of hours, I think it does better on landscapes than people. With people it's better to tell it you want a close up of the person. Then it did less cartoony and/or messed up ones. Though still some messed up faces.