The hordes of women descending into close-quarters combat over pink travel cups has made me think of other fads. Here are a few notable examples of cabbage-doll hybrids.
I’m still abusing the My Face lora at Night Cafe. This is me as Satan in Red:
Trying to create a visual pun involving a donkey-like animal with a halo (indicating that it is nice) with a pair of birds sitting on its back. 90% of the time Bing tries to draw the halo anywhere but on the donkey-like animal.
I thought this failure was funny. It placed the halo over the girl–but the halo is held in place by wires.
This is as close as I got to the image I was aiming for (even though the donkey-like animal is too big).
Trying for a spherical vacuum in a cow. This might pass, if a cow’s digestive tract was lined with tiny cow heads. (And why not, I ask you?)
Inspired by this thread, I just tried for a moose made out of chocolate chips. Note the sign.
Did you prompt it with the term mocolate choose or did it come up with that based on the prompt of chocolate chips and moose?
I literally just asked for a photo of a moose made out of chocolate chips.
Damn. I feel like that’s gotta be a coincidence.
When I try for amusing typos they are hard to get. Recently I tried making a rack of postcards for the vacation resort Pen Island that say “Welcome to Pen Island”. DE3 was both careful with the spelling and conventional with the imagery.
How do you get Bing to show an average looking woman? Someone you might meet on the street or in a 7-11. All I ever get mostly is stunningly attractive women in full makeup. I’ve used stuff like “average”, “unattractive”, “no makeup” and I still consistently get stuff like this:
If I used things like “bedraggled” etc, the clothes and hair become shabby. The woman remains stunning and wearing way too much makeup. Stop showing me supermodels!
(I’m also frustrated I can’t get that image with only the torch as a light source.)
Might be possible but good luck. You’re running up against the fact that the AI was trained on images from the internet and most images of women on the internet are (a) celebrities, (b) porn or (c) influencers and wannabe influencers off Instagram, VSCO, etc. The AI doesn’t even have a grasp on what an “average” woman looks like because the internet average ain’t the real life average and average normal women who post their photos don’t tag them as “I’m an average woman”.
With some AIs you can use a known average looking noteworthy person to prompt in that direction but Bing is allergic to known names. Best I can offer is to bust out the thesaurus and maybe expand into physical traits against the normal like “broad nose”, “round face”, “thin lips”… basically the opposite of what the AI version of woman looks like.
Thanks for the reply. Hmmm: if they can use Instagram, I’d think they could scrape a boatload of really average or not that attractive people off of Facebook.
Thanks for the suggestion. I’m trying that and now I get supermodels or cartoon characters. Damn you, Bing!
This comes close, but doesn’t really follow my prompt:
“photo of a woman with a wide nose and small eyes and thin lips and a round face, hair pulled back, in an unlit cave passage carrying a torch, walking toward the camera, dressed as an explorer. 50mm”
Definitely not broad nose or thin lips. Lips are huge. Face is roundish I guess? Didn’t ask for pigtails, wth?
Cartoon character results that also don’t follow my prompt really:
photo of a bedraggled haggard tired 30 year old woman with a wide nose and small eyes and thin lips and a round face, hair pulled back, in an unlit cave passage carrying a torch, walking toward the camera, dressed as an explorer. grainy disposable camera snapshot
Other attempts
I’d bet haggard, tired, 30yo helped a lot. I’ll experiment. Thank you.
Your other experiments were just aging that prompt I think?
Wish I had more control over the light source. Your seem to be flash photography.
Also don’t discount the utility of “grainy disposable camera snapshot”. You can also put a year or decade on the end.
I had to cheat a little but I got this:
Agreed, that does make a difference. I took out “bedraggled haggard tired”, left the age and the camera part, and got this:
More ‘average’ looking.
Then tried 15 years and got a fairly normal looking teenager:
I think your camera shot bit helped a lot. I’m unconvinced “wide nose and small eyes and thin lips and a round face” did much.
Replacing “unlit” with “dark” might have some effect.
Nah, just got 4 images that were lit with either natural light behind or overhead incandescent fixtures. Still getting credible ‘average’ humans though! (3 of 4 taking a selfie with that disposable camera.)
photo of a bedraggled haggard tired 30 year old woman with a wide nose and small eyes and thin lips and a round face, hair pulled back, in an unlit cave passage carrying a torch, walking toward the camera, dressed as a dora the explorer. grainy disposable camera snapshot.
