kAIyotei's Den

Graffiti of a punk cow with pinwheels. Two options. Pick one.

 A nice syrup drenched stack of vinyl for your hungry ears.

A thousand years ago, in the mists of the cloud forests of northern Peru, near the source of the mighty Amazon River, the mysterious shamans known as the Chachapoya Cloud Warriors reigned supreme. 

Two critters created with "asian characters" on their bodies. I added the butterfly to the chameleon image. 

Sitting in an English Garden Waiting for the Sun If the sun don't come you get a tan From standing in the English rain

A Long-Tailed Black Voodoo Toad. Created the toad in Copilot Designer, then imported the image into PhotoRoom and added the spider and some shadow and lighting effects. Finished the piece in the Pixlr app. Overall, I'm pretty pleased with the end result.

PhotoRoom makes it easy to cut out the backgrounds from multiple images, and then knit them all together in a digital collage.

 A super hero cloaked in burlap, green felt, tin foil, and cardboard. Not sure what his super powers are, but he's easy to recycle!

I saw a silly little sticker and figured it would make a good meme. The machine did a good job of making the boy look like a 70s kid. Right down to the plaid bell bottoms.

So I maxed out the number of characters for the AI this morning. Really detailed description, but it added more objects than I asked for, and didn't have a clue that "Eleodes Armata" is a stinkbug. Using the prompt: "Hopi Indian style Antique Basket Olla,    coiled wicker type baskets made of wild currant and rabbit brush. sasquatch and bigfoot and spiral designs on basket. three baskets lying in the desert near saguaro cactus. two arrowheads of turquoise and coral. deer skull with antlers in the sand. sagebrush in the background at dawn. baskets are tumbled around and decaying. a large black Eleodes armata and orange flicker feather in the bottom front of the picture with magical sun rays." All in all, it did a pretty good job with my idea. 

An punk rock band in origami style with a giant abstract tape machine in the background. Feel the noise.

Today I got a wild idea to make some Japanese kanji characters out of colored rock climbing rope. I had some other design elements already, like the Buddha and the ivory temple, so I added them to the mix and came up with this collage. The Wat Pho in Thailand is famous for its reclining Buddha, so that's where the name came from. There are also Tibetan and Chinese designs in this image.

 I was playing around with some old poetry of mine, and put a few snippets into Copilot AI. One of the phrases contained the words "Shaman Dog" which produced a shōgun coyote-like dog. I used a text to poem machine to help write the haiku.

 Saw this short history article on the origin of the Peace symbol. Interesting reading and decided to create a graphic for it. I really like the final sleepy-eyed girl that was generated. I think there is a little distortion to the fingers behind the peace sign, but not enough to make me want to fix it.  https://www.cnn.com/style/article/style-origins-peace-symbol/index.html

 I tried to recreate a scene from last Thursday. I hiked a few miles up a snowy ridge in the Wasatch Mountains and returned just at sunset. 

Sometimes when I'm reading a book, I'll stumble upon an interesting sentence and see what the AI will create with it. This sentence "It was a Russian military prosthesis, a seven-function force-feedback manipulator, cased in grubby pink plastic" was from Necromancer by William Gibson. It was describing the prosthetic arm of the bartender Ratz in the Chatsubo Bar. So here is what the machine made from it:

 I'm not sure what I expected when I asked for interlocking fish, snakes, and lizards, but here it is. I find it kind of soothing.

My first shot at creating a fantasy image. I combined two different images into this scene. I think the AI did a great job on the colored stones in the foreground. It took a lot of tries to get the wizard and the elk just right. 

 I'm not sure why AI mangles text so badly. I mean I type in exactly what I want and it scrambles it or adds extra letters, or just ignores a word completely. I'm guessing this is a copyright thing? Anyway, here's a design I used for a post about one of my guilty pleasures... The song "You Sexy Thing" by Hot Chocolate. And it almost got it right text-wise, missing it by just letter and spelling "Sold Gold" instead of' "Solid Gold" which I actually like and chalk it up to a so-called Happy Accident.

 Been testing out a new AI called Stable Cascade which has an online demo ( they call a playground). I like the fact that one of options allows you to select a portrait or landscape images, instead of just a square shape. The first few I tried really turned out good, high rez and not wonky, like this one: But then the next ones were not so good. Horses with three legs, or blurry incomplete horse and rider. I am still amazed by how good the programs are getting, and how quickly! If I were making my living as a graphic designer, I'd be very nervous.

I cheated a bit on this one. Saw a picture of a fox that I liked while exploring, so I copied the prompt and changed the word fox to coyote. "A realistic image of a coyote, walking on a dirt path through a forest with tall trees and colorful leaves, with rays of sunlight and mist creating a magical atmosphere, morning wildlife photography." I'm using Copilot Designer, which is so far my favorite AI machine for making art. It it just constantly excellent and I don't have to wait for 10, 20, or 30 iterations to get one I like.

I called this  Experimental Comic design "Gangurru tjukurrpa" which means Kangaroo Dreamtime in Australian Aborigine language. It was made using several images that I put together using the Pixlr app for collages.   

 Today I was playing around making faux postage (fake stamps) AKA artistamps. I hard a hard time getting the AI to give me stamps with all four perforated corners. It kept chopping off at least one side. So I tried again and again ... finally ending up with a lot of images. I decided to put some of them together in a digital collage. I also made a model for NightCafé and got a good image on the third try. The man with the green guitar. I added another image next to him, but I didn't like it much. So I created this girl in red to fill out the collage. I used the PhotoRoom app to put it all together.   And here are some of the incomplete stamp images. If anyone knows how to do it right, give me a shout. I used Bing Image Creator to make them.

NightCafé Creator is one of the other AI machines that I use. It isn't free but it isn't too expensive either.  One thing I really like about it is the ability to create a model of yourself using photos. After "training" it, I think it did a pretty good job. Maybe my hair and beard are a bit longer, but that's cool with me. One thing I don't like is how it often adds extra objects, mangles my arms, legs, and twists fingers, and sometimes adds six fingers to my hand.  Check out these pix and you can see what I mean. Although, it did make me into an Excellent Pirate:

 While testing out a still life composition using AI, the machine decided to light one of my paint brushes on fire. I liked it so much I asked for them all to be aflame. I called it "Still Life without Cabbage" for some strange reason.

Billions of cicadas will emerge in the U.S. this year in a rare double-brood event.  This year's dual emergence is a rare, synchronized event.  This spring, two different broods of cicadas — one that lives on a 13-year cycle and the other that lives on a 17-year cycle — will emerge at the same time. That dual event last occurred in 1803. What will we do with them all? May I suggest Cicada Parmigiana?

Happy Chinese New Year. This is the year of the dragon. 

Okay, a bit mushy I suppose. But i was thinking of the great Sir Doug and the Texas Tornados song... Give Back the Key to My Heart.  On YouTube here:

This is a place to drop my AI creations, record the wacky thoughts I have while creating them, and discuss the process involved.