![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
So the "forexposure_txt quotes turned into comics with Midjourney art" link from the last post is gone :(
Good news: Archive.org has a saved version! Bad news: it has a wild "getting stuck in an infinite loop of reloading" problem. Saving-grace news: it looks like if you hit Stop at just the right time, you can get a version where...most of the images have loaded.
Gonna quietly save the whole thing to my own hard drive now. Just for posterity.
It was originally posted back in August, btw. Some of this AI-art discussion makes it sound like it's a revolutionary new concept that just launched this month, and, no? This isn't the first iteration of AI art. Probably not the tenth, or even the twentieth. People have been refining it for years.
Have a post from August 2018, with sample prompts and their AI-generated images. It's fascinating -- you can tell the software associates the keywords with certain shapes, or patterns, or blobby impressions, but can't put all those concepts together in a comprehensible way.
And a post from July 2021, trying to get a picture of sheep grazing on a green hillside, but, come on, a good picture, please? Lots of iterations, comparing the results when you add different keywords. Some of the scenery comes out really lovely! The sheep...dlet's just say they never quite look like sheep.
...And this is a 2022 post with an AI's best attempt at a (simple, 4-panel) comic.
Fellow comic artists: I, uh, don't think we have to worry about the computers taking our jobs any time soon.
Good news: Archive.org has a saved version! Bad news: it has a wild "getting stuck in an infinite loop of reloading" problem. Saving-grace news: it looks like if you hit Stop at just the right time, you can get a version where...most of the images have loaded.
Gonna quietly save the whole thing to my own hard drive now. Just for posterity.
It was originally posted back in August, btw. Some of this AI-art discussion makes it sound like it's a revolutionary new concept that just launched this month, and, no? This isn't the first iteration of AI art. Probably not the tenth, or even the twentieth. People have been refining it for years.
Have a post from August 2018, with sample prompts and their AI-generated images. It's fascinating -- you can tell the software associates the keywords with certain shapes, or patterns, or blobby impressions, but can't put all those concepts together in a comprehensible way.
And a post from July 2021, trying to get a picture of sheep grazing on a green hillside, but, come on, a good picture, please? Lots of iterations, comparing the results when you add different keywords. Some of the scenery comes out really lovely! The sheep...dlet's just say they never quite look like sheep.
...And this is a 2022 post with an AI's best attempt at a (simple, 4-panel) comic.
Fellow comic artists: I, uh, don't think we have to worry about the computers taking our jobs any time soon.
