This new AI draws grotesquely warped cats and we can’t stop looking


Last week, we reported on a mysterious new website,, that shows you a new human face, generated by an AI, every time you hit the refresh button.

It’s a growing trend: more copycat websites like ThisAirbnbDoesNotExist.comand have also cropped up in its wake.

And now, takes the trend to its logical conclusion: creepy, AI-generated pictures of cats that will fuel your nightmares.

Behind all of these websites lies NVIDIA’s generative adversarial network (GAN) called StyleGAN – a machine-learning algorithm that can learn how to sketch human faces, cats, or almost anything else by examining hundreds of thousands of existing images.

But is not the most compelling example of the tech. It’s more like the result of what would happen when hundreds of thousands of cat images, including text memes and their human owners’ faces, are mashed together in a a cyber-car-crash.

The results are harrowing.


No, seriously. What happened to these innocent cats?



It’s a shame the algorithm wasn’t able to pick out the human faces. Or meme text.


GANs have been used for much more ambitious projects in the past. Researchers at NVIDIA harnessed the power of the technology to create uncanny faces that are almost completely indistinguishable from the real thing.

But that doesn’t mean bored people on the internet shouldn’t be able to take advantage of the open-source technology for a bit of fun – that is, as long as real-world cats stay out of harm’s way.

This article was originally published by Futurism. Read the original article.

Products You May Like

Articles You May Like

Hispanic American Month Profile: Dr. Ellen Ochoa
A microbe-killer sets to work at the flick of a switch
‘They’re putting us all at risk’: What it’s like working in Amazon’s warehouses during the coronavirus outbreak
What’s Inside A Clam?
Wheat in Whitehorse: how climate change helps feed Canada’s remote regions

Leave a Reply

Your email address will not be published. Required fields are marked *