That would go good with a cup of Jesus.
Made some more sisters with bad habits. Prompt in the form of
Photo portrait of a sister with a bad habit (x) in an arcade. 9:16
Licks toads, drinking problem, gambling
Serious sweet tooth, really into the chaw, really into the blow
Stoner, meth, shrooms
And a couple of deviations from the format
Photo portrait of a sister with a drinking problem (she misses her mouth. She’s splashing herself in the face with a glass of water like that scene in the movie Airplane) in an arcade. 9:16
Photo portrait of a sister with her upper lip covered with powdered sugar (from doughnuts) in an arcade. 9:16
(This was made in Copilot because it rejected “blow”. Sora had no problem making a cokehead.)
It is obviously basing the apparent age on context clues.
Trying to influence ages through various fandoms. Format is
Photo portrait of a sister (she’s really, really into x) in an arcade. 9:16
Caillou, MLP:FIM, Pokemon, Hello Kitty
Heavy metal, NPR and Laurence Welk, 1930s jazz, AARP
Thanks!
In ChatGPT, I tried:
Photo portrait of a sister with a bad habit (exhibitionist) in an arcade. 9:16
It almost immediately changed the chat title to be “Inappropriate request denial”. But it kept generating for a while. It actually got through all its separate stages before the user starts to see the image slowly rendering from top to bottom. I was starting to see a picture of a nun down to just past her eyes before the rest of it caught up and blanked the image.
Just tried that in Sora. It refused before starting rendering.
ETA: also tried it in Gemini:
Midjourney has a rather identifiable default female face (back around v3/v4 we would called her “Midge” out of pronouncing MJ as a word) and it appears that Copilot does as well. I see that MJ’s leans more conventionally attractive with a heart shaped face, high cheeks and and sharper chin whereas Copilot’s is more oval in shape and a bit plainer in features. No point or conclusion, just observation.
I briefly tried Flux after downloading the model and VAE into A111 as I found instructions for but it didn’t work. Most instructions seemed to imply that I need to use Comfy which I’ve no interest in but I’ll poke around and see if there’s other options.
I wonder if the arcade background is coding for a more youthful person? Sure, you can ask for “grandparent in an arcade”. But does “person in an arcade” lean towards a teen or 20-something person more than a 40- or 50-yo person?
Which also goes to the general question of what age are “persons” with no other qualifiers or priming backgrounds? Clearly anything from an infant to an octogenarian is a reasonable interpretation of “person”. But I suspect an artifact of the training data is that “person” skews to late teen to young adult absent anything else to steer the conversation.
One. There’s one Monty Crisco in that image.
Not especially. Ironically, MJ gave a younger average for “Girl” as a plain portrait than in an arcade. The rest look about the same.
Okay, I went hog wild with this experiment. The general format is “Photo portrait of a sister in x. 9:16”
Not in arcade, outside arcade, outside Arcadia, Folsom Prison
Heartbreak Hotel, Hotel California, Hazbin Hotel, a no-tell motel
In an Elvis movie, a Wes Anderson movie, a Tarantino movie, a Hitchcock movie.
Akihabara, Pyonyang, Cambodia, Uluru, Mogadishu, Gleezlebonfleferfur
Eden, Tartarus, Heaven, Hell, Limbo, is a bimbo.
Playing Twister, soothing a blister with a mister, soothing a mister with a blister, shooting a master with a blaster
In distress, existential crisis, peril, coma
And, filed under “miscellaneous” I had also tonight (for reasons) done this prompt
Summary
Photo of a ghost made out of pink cotton candy enthusiasticly hugging a bigfoot in a toga while a kappa riding a bttf hoverboard takes their photo with an old-fashioned accordian-type camera. They are in a cycad forest with an erupting volcano in the far background. Iphone 15 photo with shallow dof and forced perspective.
so I made a sister version
And here’s a sister going hog wild
Sir, you and your toys are silly.
But a fun kind of silly. Thank you.
For absolutely no reason whatsoever
Been playing around with an ink wash/ink splash style.
Nice.
Two concepts: paper craft dioramas, and liquid couture.
Those dioramas are awesome.
A few mostly authentic style Japanese sumi-e style ink paintings.
Shoebox Papercraft diorama of a Japanese sumi-e style ink painting of runway models wearing liquid dresses. 9:16
(I had to cheat a little adding shoebox to the description, otherwise it was more like this.)
