rururuku-scribbles

[rururuku-scribbles]

Gift art, of a sort, for @reachartwork , who's been getting some absolutely horrible hate messages recently. I scrambled and ciphered anon's screed and then turned the resulting nonsense text into a prompt. This grim grinner and the bluegreen Circle of Protection Against Haters were the most interesting things to bubble up from the chaos soup.

Rage Against the Machines - AI imagery by Rukuruku via SDXL

These images are raw AI output without significant iteration or other human intervention, and therefore public domain.

[rururuku-scribbles]

Nobody Cares Like a Bar-Bear-ian

I asked Bing for 'Care Bears and/or Care Bear Cousins in the style of a fantasy TTRPG monster manual entry'. What it gave me was a lot closer to 'Care Bears as a playable species option for a fantasy TTRPG', but I can't say I'm disappointed with the result even if it wasn't what I was trying to generate. I'm actually genuinely interested now in what their statblock would be like.

Sensursa Characters- AI imagery by Rururuku via DALL-E 3

These images are raw AI output without significant iteration or other human intervention, and therefore public domain.

[rururuku-scribbles]

An AI AU Almost Advent Calendar

Or should that be 'Advent Children' calendar? Aka, personal project which I noodled with over at PixAI off and on over most of December.

This is the first in what will probably be an occasionally recurring series, so if the pic above the cut is not to your liking, you can either block me or blacklist the tag 'the ongoing adventures of deburos' and this is the last you'll see of it.

So, short explanation: These are what ultimately resulted of a daydreamed-about but never actually published fanfic wherin Sephiroth goes AWOL from Shinra and, deciding that he never wants to be their- or anyone else's- tool again, sets about molding himself into less of a weapon.

(And if you're an MCU fan and the general outlines of that plot structure sound vaguely familiar to you, I will admit to having been inspired by a Stucky fic I read some time ago)

The issue I ran into- aside from, you know, the usual (if you're a writer, esp. a fanfic writer, you know) was that the parts that I was most interested writing in weren't any of the actual plot or character development beats. Or, for that matter, any of the stuff that would appeal to people looking to read about Sephiroth putting on weight with a kinky framing. What I most wanted was what would have amounted to the epilogue- chubby!Seph just quietly going about living his best life.

And so, having managed to achieve a result that scratched that itch while tooling around with PixAI's Moonbeam model, I kind of just gave myself the permission to skip to the good parts. These were kind of intended as a sort of advent calendar, but I didn't quite come up with 25 seperate ideas (though I came pretty close), and events in meatspace meant that they didn't exactly get posted on a one-per-day schedule, either. Still, I'm pretty pleased with myself about these, even if I had some social anxiety over actually posting them where other people can see and comment.

Some of these are direct, raw AI output, while others involved significant amount of curation, tweaking, and img2img iteration back and forth across multiple websites/tools- primarily PixAI (where the final pieces were produced) and Playground, but a little bit of Civitai is mixed in as well.

If you've scrolled this far, I hope that your holidays were happy and that you have a great 2024.

[rururuku-scribbles]

The Proof is In the Prompting

So, remember what I said in my previous post about how the images in question illustrated something about the impact things do (and don't) have in an image prompt? Here's an explanation of what I was talking about with that. Plenty of other people have talked about how some of the common negative prompting that people enter ("bad anatomy, extra fingers, poorly drawn face, ugly, distorted," etc) involves a significant amount of magical thinking- that, because of how most widely available ai art generators are built on top of the snapshot-of-the-whole-internet Common Crawl data and the way those images were labeled, telling the AI to exclude bad hand anatomy or whatnot doesn't reliably get it to not generate wonky hands. The same goes for some common positive prompting, like "trending on artstation".

But, what if I told you that prompting the AI to draw in the style of a living artist is also often just wishful thinking? The image on the left has 'by greg rutkowski' as part of its prompt- Playground automatically appends that (as well as 'trending on artstation, sharp focus, studio photo, intricate details, highly detailed') to the end of your prompt when you create an image using the 'Delicate Detail' style. The image on the right was generated from the same seed, same prompt (including the 'trending on artstation, sharp focus...' part), with the only thing changed being that it was created with no style applied and with 'by greg rutkowski' removed. They're not identical, but they're close enough to be recognizably variations of the same picture, and if they're not in exactly the same style, they're very close.

As a control, this pair of images are from me accidentally running the exact same seed and prompt twice- there's a tiny bit of variation, but it's easy to miss if you're not looking for it and depending on your visual acuity and the size of the screen you're seeing this on, it might not be visible even if you are looking.

And here's the reason why I ended up accidentally duplicating a generation- the image on the right is where I accidentally deleted my negative prompts. Still has the 'by greg rutkowski', though! You can still see that it's a variation of the same picture (/generated from the same seed), but the difference is much greater than with the presence or absence of prompting for a named artist. (Very possibly it could be considered a stylistic difference, but I don't have the technical vocabulary/art school background to be confident about that.)

So, what's the point here? What's the takeaway, what did I learn from these mis-gens? First, naming an artist in a prompt doesn't necessarily affect the end results all that much. For very widely reproduced/riffed on/repeatedly included artists like Van Gogh, Caravaggio, or any of the other 'Masters' and big influential names you get taught about in art classes, sure, you get a demonstrable difference, but while Greg Rutkowski has potentially valid reason to be upset about people prompting AIs using his name, it doesn't actually do as much as the people prompting probably were thinking it does. Without directly going image-to-image or being specifically trained on a particular artist's works, AI (or at least, Stable Diffusion/SDXL and its descendants) isn't capable of replicating the style of most random living artists, because it doesn't 'know' what their style is in the first place.

Second- those negative prompts may not do exactly what they say that they're doing, but they sure are doing something. Excluding them there may have been an accident, but I might want to noodle around with doing it on purpose sometimes just to see what happens.

[rururuku-scribbles]

Non-Plasticine Ladies

More ladies loving ladies within the enchanted forest- some of the not-so-Playdoh results from the same prompt as the previous post.

Incidentally, also an illustration of what does and doesn't matter in an image prompt and how that may not be what you think- but that's a discussion for another post.

The Nymph and Her Fairy Protector - AI imagery by Rukuruku via Stable Diffusion 1.5

These images are raw AI output without significant iteration or other human intervention, and therefore public domain.

[rururuku-scribbles]

(AI imagery by @renaissanceofthearts via Midjourney + post-generation editing)

Island...

[rururuku-scribbles]

Playdoh Lesbians Embrace in the Enchanted Forest - AI imagery by Rukuruku via Stable Diffusion 1.5

These images are raw AI output without significant iteration or other human intervention, and therefore public domain.