An artificial intelligence that can create photorealistic images of people that don't exist was recently let loose on cats - without quite the same level of success.
StyleGAN, a type of algorithm called a neural network, made headlines in December when it was fed thousands of photos of real human faces, and spat out images of fake humans indistinguishable from the real thing.
But the same algorithm, fed thousands of images of cats, created monsters.
Janelle Shane, a research scientist from Colorado, shared many of the horrors on her blog aiweirdness.com.
She said the algorithm used images pulled from photosharing site Flickr.
"Thanks to that big dataset, a new method of generating images, and a staggering amount of computer power, its human faces are indeed impressive.
"But, to prove that their method doesn't just work for human faces, they also generated bedrooms, and cars… and cats. The cats are so much fun."
The reason many of the 100,000 cat images often came out looking cat-astrophic was that cats "can be in so many different poses", compared to portraits.
"I've noticed that for some reason, whenever it generates kittens, one is normal and the rest are haunted. There must be something difficult about pictures containing multiple subjects."
Some of the images came back with gibberish text on them, thanks to the popular internet 'lolcat' trend. Many looked like Tardar Sauce, a cat that found internet fame in 2012 as 'Grumpy Cat'.
"She's only one cat, but she had a profound impact on what StyleGAN thinks cats look like."
The code, and thousands more images of cats that don't - and probably shouldn't - exist are available online.
Newshub.
