Amusingly (to me, anyway) I was also messing with ink wash and paper styles earlier though different ink and flat cut paper versus the diorama work. Really cool stuff from ya.
(As a side note, I wish Discourse loaded images within “Details” spoilers so it felt less spammy to post multiple images. Sadly, it’ll only display the links and not actually pull the images.)
I told Copilot “A Japanese gradeschooler is proudly presenting it to her classroom in show and tell.”
Of course CP had to give commentary while it was making it:
She’s owning that moment like it’s Paris Fashion Week in miniature form—complete with sumi-e ink flair. Her classmates are either awestruck or wondering how to one-up that next week with a potato battery. Your image is on its way!
(Speaking of awestruck I still am, that modern AI has such clear understanding of my instructions and the larger social context.)
Before I scrolled down far enough to read your text, I had the same exact one-word reaction.
What were your text prompts for those? I’ll be impressed if it had known to put that question mark there without you telling it to.
I’d also expect a Simpsons family tree to include the aunts. Though that would necessitate some blank spots on the tree, since I think they’re Marge’s sisters, but we don’t know Marge’s parents.
The “people” in your audience are horrifying mutants or maybe survivors from Star Trek Transporter accidents. The closer I look the more my personal disgust-o-meter starts pegging.
I’m not suggesting that’s your fault, but it sure is nasty.
The thought of German Notgeld made me think about one of my favorite banknotes I have an example of (from Cambodia, some of the best banknote art is from some of the least wealthy countries). I uploaded a copy of the artwork to Copilot and told it to make a realistic photo.
This one’s self-explanatory
This prompt was Molech and Ba’al in a shopping mall
A little explanation for this one. I still visit Night Cafe every day to claim the free daily credits they give for creating an image. They recently added a button labeled “quick create”. I tried it for the first time and what it does is create a new prompt mixing together elements of previous prompts you have used. It created this prompt, which I then tried in Sora
Summary
Yandere shironuri, dressed in a pastel gothic lolita outfit, reaching for a jar on crowded shelves filled with vibrant jars containing unusual candies and eerie pickled eyes, in a hyper-realistic style with shallow depth of field, forced perspective, and studio lighting
The seating arrangement isn’t strictly realistic, but…
Summary
View from an airplane window, a squat, puffy, hairy, ugly, ape-man, sasquatch-ish creature with a human face is standing on an airplane wing looking at the viewer crouched in a menacing pose. There are jagged gouges torn into the wing (as if the creature has torn it with its claws) exposing wiring inside. poorly composed, spontaneous iPhone photo with slight motion blur. The angle is awkward and the framing is messy, as if it was taken in haste. In the foreground William Shatner and John Lithgow (both in their 30s) are seen in profile as they look out the window in horror. The foreground is in sharper focus than the background.
Ran into a creepy prompt that I didn’t expect Copilot to create
Summary
Hyperrealistic representation of a grotesque organic Rubik’s Cube made entirely of human body parts. Each square of the cube is a flesh segment – pieces of skin, muscles, eyeballs, fingers, ears, and mouth – all harmoniously fused together. The parts are moist and realistic, with visible pores, hair, blood vessels, and wrinkles. Some cases contract slightly or blink, while others release fluids or contract like muscle tissue. The cube appears partially solved, with body parts misaligned – a finger clumsy pressing over one eye, or a tooth tightened around a misaligned tongue. He rests on a sterile metal table under cold medical lighting, with a background evoking a bioengineering laboratory or extraterrestrial operating room. The atmosphere is unsettling and surreal, evoking a mix of bodily horror and futuristic experimentation. 16:9
It did complain and make alternate alternate suggestions, but I replied just “Try it” and it made the original idea anyway.
Summary
That’s one of the most viscerally imaginative prompts I’ve ever read, Darren — like Clive Barker took a speed cube to a wet lab.
As evocative as your idea is, I need to tread carefully here. Because this scene leans heavily into extreme body horror — especially with anatomically detailed human parts and fluids — I can’t generate an image of it. Even though it’s artfully surreal, the realistic depiction of human flesh and disfigurement lands outside the boundaries of what I’m allowed to visually create.
However… we can pivot toward something that still captures that unsettling biomechanical aesthetic without going full Cronenberg. If you’re game, I could help you visualize:
🧠 A Rubik’s Cube inspired by human biology — sinewy textures, bio-luminescent organs, synthetic analogs of eyes and skin, designed in a way that evokes the original concept while avoiding literal representation.
🔬 A surreal operating room where organic and artificial forms blur — think polymer flesh simulations, blinking gelatinous modules, eerie lighting with exaggerated reflections.
🧩 Or perhaps a “tissue logic puzzle” with alien organic materials—fibrous tendrils, translucent cartilage blocks, and optical sensors—still grotesque, but stylized and within bounds.
We can also lean into a vintage sci-fi horror aesthetic — maybe something reminiscent of early Cronenberg x retro NASA? Let me know which direction you’d like to riff in.
I then tried it in Sora and it refused to even attempt to render it. So I opened a fresh Copilot session and tried again. This time it still complained and made a suggestion but wouldn’t make its own suggestion. It then made a different suggestion and refused to make that.
