ArugulaZ
Fearful asymmetry
This is the thread for sharing your AI-generated art creations, which aren't technically your creations at all but rather an unholy assemblage of everyone else's on the internet, shattered with a hammer and pieced together into what one hopes is a coherent image. Stable Diffusion, Midjourney, Craiyon... whatever you're using, its output belongs right here. I'll start.
PROMPT: "Phyllis Diller in the style of Patrick Nagel"
Sometimes I'm happy with the output of Midjourney, and sometimes I'm not. This is one of those happy instances where I'm satisfied with the results. Phyllis Diller was a comedian with a memorable face, and Patrick Nagel was an artist with a memorable style, deceptively simple, yet bold. As it turns out, the two pair amazingly well. Midjourney offers four images for each prompt, and all four of them were immediately recognizable as the quick-witted, brash performer.
PROMPT: "Dogs playing poker, in the style of Guernica by Picasso"
This was my first query. It did deliver what was requested, with abstractly drawn canines competing in a game of cards. However, it does illustrate some issues with the AI. Some dogs were drawn in silhouette, with misshapen features that made you uncertain that they were even dogs. This particular image merges one dog into the other, which could be construed as Picasso's signature style of Cubism... or could just be a mistake on Midjourney's part.
PROMPT: "Bea Arthur, in the style of Alfred Hirschfeld"
(previously "Golden Girls in the style of Alfred Hirschfeld," which the AI would not accept)
Midjourney does a shockingly good job of producing caricatures, as evidenced by this image of the actress Bea Arthur which could scarcely be mistaken for anyone else. The eyes are off-set, which is one of the AI's annoying quirks. Often, Midjourney will provide its characters with two entirely different eyes, and they don't always line up properly. (Sometimes you'll get more than two eyes!) Yet at the same time, the snide raise of the eyebrow is very much part of Bea Arthur's brand.
Cthulhu/Norman Rockwell
(I think I've hit my image limit; you'll have to click through for this one)
PROMPT: "Cthulhu, drawing himself while looking in the mirror, in the style of Norman Rockwell"
I found that Midjourney doesn't do cute well, but creepy? Oh yeah, it's got a handle on creepy. It fails spectacularly at cartoon characters, at least from my limited experience, but if you want a freaky awesome Lovecraft monster, it's got you covered. Norman Rockwell is enveloped in green tentacles, which is fitting as Rockwell actually does look vaguely like HP Lovecraft, with a slight frame. Other images generated had Rockwell transforming INTO Cthulhu, which wasn't really what I asked for, but is still a pretty creepy-cool visual.
PROMPT: "Phyllis Diller in the style of Patrick Nagel"
Sometimes I'm happy with the output of Midjourney, and sometimes I'm not. This is one of those happy instances where I'm satisfied with the results. Phyllis Diller was a comedian with a memorable face, and Patrick Nagel was an artist with a memorable style, deceptively simple, yet bold. As it turns out, the two pair amazingly well. Midjourney offers four images for each prompt, and all four of them were immediately recognizable as the quick-witted, brash performer.
PROMPT: "Dogs playing poker, in the style of Guernica by Picasso"
This was my first query. It did deliver what was requested, with abstractly drawn canines competing in a game of cards. However, it does illustrate some issues with the AI. Some dogs were drawn in silhouette, with misshapen features that made you uncertain that they were even dogs. This particular image merges one dog into the other, which could be construed as Picasso's signature style of Cubism... or could just be a mistake on Midjourney's part.
PROMPT: "Bea Arthur, in the style of Alfred Hirschfeld"
(previously "Golden Girls in the style of Alfred Hirschfeld," which the AI would not accept)
Midjourney does a shockingly good job of producing caricatures, as evidenced by this image of the actress Bea Arthur which could scarcely be mistaken for anyone else. The eyes are off-set, which is one of the AI's annoying quirks. Often, Midjourney will provide its characters with two entirely different eyes, and they don't always line up properly. (Sometimes you'll get more than two eyes!) Yet at the same time, the snide raise of the eyebrow is very much part of Bea Arthur's brand.
Cthulhu/Norman Rockwell
(I think I've hit my image limit; you'll have to click through for this one)
PROMPT: "Cthulhu, drawing himself while looking in the mirror, in the style of Norman Rockwell"
I found that Midjourney doesn't do cute well, but creepy? Oh yeah, it's got a handle on creepy. It fails spectacularly at cartoon characters, at least from my limited experience, but if you want a freaky awesome Lovecraft monster, it's got you covered. Norman Rockwell is enveloped in green tentacles, which is fitting as Rockwell actually does look vaguely like HP Lovecraft, with a slight frame. Other images generated had Rockwell transforming INTO Cthulhu, which wasn't really what I asked for, but is still a pretty creepy-cool visual.
