AI-generated images thread

Argent Stonecutter

Emergency Mustelid Hologram
Joined
Sep 20, 2018
Messages
6,507
Location
Coonspiracy Central, Noonkkot
SL Rez
2005
Joined SLU
Sep 2009
SLU Posts
20780
Half the price, hu? You can just download the model for free.
Free as in puppies.

Edit: The cost is not in actual money. I get 5 tokens every day I remember to check in, and SDXL costs me a token, and Lightning costs me 0.5 tokens. It can take me five or six tries to get a good picture of a Dingo serenading Darryl Hannah dressed as a replicant or whatever, so having twice as many tries per day is handy.
 
Last edited:

Free

A wink and a smile.
VVO Supporter 🍦🎈👾❤
Joined
Sep 22, 2018
Messages
37,163
Location
Moonbase Caligula
SL Rez
2008
Joined SLU
2009
SLU Posts
55565

Bartholomew Gallacher

Well-known member
Joined
Sep 26, 2018
Messages
6,101
SL Rez
2002
Time for another round of hideous images created using Dreamshaper XL Lightning, all rendered natively in 1024x1024 with 4 steps only, here we go...

Donald Trump is the poop train - it really shows his love for poop...

A fusion of Walt Disney with the Donald.

Trump the dictator

Some random anime girl to relief your pained eyes.



Chubby mullet Trump.

A young chinese girl in a workshop.
 
  • 1Like
Reactions: Govi

Free

A wink and a smile.
VVO Supporter 🍦🎈👾❤
Joined
Sep 22, 2018
Messages
37,163
Location
Moonbase Caligula
SL Rez
2008
Joined SLU
2009
SLU Posts
55565
Have you ever seen an Asian person with a white person, whether that’s a mixed-race couple or two friends of different races? Seems pretty common to me — I have lots of white friends!

To Meta’s AI-powered image generator, apparently this is impossible to imagine. I tried dozens of times to create an image using prompts like “Asian man and Caucasian friend,” “Asian man and white wife,” and “Asian woman and Caucasian husband.” Only once was Meta’s image generator able to return an accurate image featuring the races I specified.

Hmm. Meanwhile, it looks like Meta is trying to fix the problem by temporarily turning off Asians - so to speak.

 
  • 1Facepalm
Reactions: Govi

Cristiano

Cosmos Betraying Fiend
Admin
Joined
Sep 19, 2018
Messages
5,003
SL Rez
2002
Joined SLU
Nov 2003
SLU Posts
35836



Hmm. Meanwhile, it looks like Meta is trying to fix the problem by temporarily turning off Asians - so to speak.

On a related note, in various image generators, namely Dall-e 3, a prompt with handsome man or men very rarely ever produces a black or asian man. I have generated hundreds of images and have to specifically prompt for black or asian.
 
  • 1Facepalm
Reactions: Govi

Argent Stonecutter

Emergency Mustelid Hologram
Joined
Sep 20, 2018
Messages
6,507
Location
Coonspiracy Central, Noonkkot
SL Rez
2005
Joined SLU
Sep 2009
SLU Posts
20780
I just spent 20 minutes and a dozen previews trying to generate of a picture of a dingo and an orangutan, only to get an endless sequence of red-furred dingoes with varying numbers of spidery extra limbs. Two different programs. generative algorithms do an awful job with descriptions of more than a single subject.
 

Free

A wink and a smile.
VVO Supporter 🍦🎈👾❤
Joined
Sep 22, 2018
Messages
37,163
Location
Moonbase Caligula
SL Rez
2008
Joined SLU
2009
SLU Posts
55565
Card game developer says it paid an 'AI artist' $90,000 to generate card art because 'no one comes close to the quality he delivers'
The maker of digital trading card game Champions of Otherworldly Magic says it has spent $90,000 on card art, the entirety of which has been paid to a single "AI artist" who receives $15,000 per month despite dedicating less than two full work days to the project each month.

"We pay our AI artist 15,000 USD per month for exactly 10 hours of work," reads an X post from the official Champions TCG account. "Why? In that time, he still makes HUNDREDS of AMAZING bits of artwork—ASTRONOMICALLY FASTER than ANY team of traditional artists.
"His art is 100% AI generated, yet it has no extra fingers, no generic designs, no mistakes... It has consistent evolutions, skins, alt art styles—literally no one is on his level. We don't care how he makes it, we only care that the end user enjoys our game."
The work is more AI-assisted.
The card images aren't quite "100%" AI generated, as the X post says. Malec says they're also touched up by hand: "AI can do bulk of the work/initial generating but to make sure no errors, extra fingers, etc everything needs to be edited and filtered."
 
  • 1Flower
Reactions: Govi

Free

A wink and a smile.
VVO Supporter 🍦🎈👾❤
Joined
Sep 22, 2018
Messages
37,163
Location
Moonbase Caligula
SL Rez
2008
Joined SLU
2009
SLU Posts
55565

More (as in various AI tools have similar problems) on the topic.

 

Cristiano

Cosmos Betraying Fiend
Admin
Joined
Sep 19, 2018
Messages
5,003
SL Rez
2002
Joined SLU
Nov 2003
SLU Posts
35836

Free

A wink and a smile.
VVO Supporter 🍦🎈👾❤
Joined
Sep 22, 2018
Messages
37,163
Location
Moonbase Caligula
SL Rez
2008
Joined SLU
2009
SLU Posts
55565
Goodnight, DALL-E 2. You shall be missed. Maybe.

Last week, OpenAI turned off the ability for new customers to purchase generation credits for the web version of DALL-E 2, effectively killing it. From a technological point of view, it's not too surprising that OpenAI recently began winding down support for the service. The 2-year-old image generation model was groundbreaking for its time, but it has since been surpassed by DALL-E 3's higher level of detail, and OpenAI has recently begun rolling out DALL-E 3 editing capabilities.

But for a tight-knit group of artists and tech enthusiasts who were there at the start of DALL-E 2, the service's sunset marks the bittersweet end of a period where AI technology briefly felt like a magical portal to boundless creativity.
 

Noodles

The sequel will probably be better.
Joined
Sep 20, 2018
Messages
4,817
Location
Illinois
SL Rez
2006
Joined SLU
04-28-2010
SLU Posts
6947
I just spent 20 minutes and a dozen previews trying to generate of a picture of a dingo and an orangutan, only to get an endless sequence of red-furred dingoes with varying numbers of spidery extra limbs. Two different programs. generative algorithms do an awful job with descriptions of more than a single subject.
I could almost never get any AI system to properly generate an "and" or "with". It would always result in, at best, two of one of the subjects. Or sometimes an unholy fusion.
 
  • 1Agree
Reactions: Argent Stonecutter

Cristiano

Cosmos Betraying Fiend
Admin
Joined
Sep 19, 2018
Messages
5,003
SL Rez
2002
Joined SLU
Nov 2003
SLU Posts
35836
I just spent 20 minutes and a dozen previews trying to generate of a picture of a dingo and an orangutan, only to get an endless sequence of red-furred dingoes with varying numbers of spidery extra limbs. Two different programs. generative algorithms do an awful job with descriptions of more than a single subject.
This is stable diffusion, simple prompt "dingo and orangutan playing in a field". This is what I got:

621
 
Last edited:

Argent Stonecutter

Emergency Mustelid Hologram
Joined
Sep 20, 2018
Messages
6,507
Location
Coonspiracy Central, Noonkkot
SL Rez
2005
Joined SLU
Sep 2009
SLU Posts
20780
OK, can you get it to make a picture of a numbat that looks like a numbat?

Ubuntu 24.04 Noble Numbat is coming out this month, so I wanted to generate new wallpaper for it.

A numbat is a marsupial carnivore that eats ants. Like most marsupial carnivores, sich as the Tasmanian Tiger, it has lateral stripes across its butt.



GANs and other generative automation almost uniformly stripe it longitudinally, like a chipmunk.

They also make it rounder with a rounded face like a chipmunk too.



I managed to get lateral stripes out of Dall-E but it was striped all over its body like a Zebra. And still round like a rodent.

 
  • 1Interesting
Reactions: Govi