art

Disconnect
Pitch work is basically when a director, writer, producer, or any combination of those get together with an artist and say, “We want to pitch to studios and we need imagery.” All of that has now been given to generative AI.
Fascinating interview with concept artist Karla Ortiz about the impact of generative AI on her industry.
stablediffusionlitigation.com
"At min­i­mum, Sta­ble Dif­fu­sion’s abil­ity to flood the mar­ket with an essen­tially unlim­ited num­ber of infring­ing images will inflict per­ma­nent dam­age on the mar­ket for art and artists."
Describing image models as sophisticated collage tools takes some of the mystery out of AI and makes it clear work is being used without consent. This essay has a clear description of the diffusion process.
YouTube
I'm so tired of hearing about crypto scams. This video takes the time to explain why that world of NFTs and crypto is such a wretched hive of scum and villainy. This explainer is well done and worth the time and I hope it can help our collective consciousness move on.

See also: David Rosenthal’s EE380 Talk
A record on a turntable with digitally altered colors
Summer Breeze
Photo of a leaf under running water with shifted colors
The Leaf Aquatic
A red flower with digitally altered colors
Red Flower #everyfilter
a picture of a factory colorized with gradients
industrial gradients
an old wind-up robot toy on a colorful digitally altered background
retro robot
a photo of daisy with digitally altered colors
daisy impression
colorful digitally altered photo of a sunflower
sunflower impression

VQGAN + CLIP + Unreal Engine + Surrealism

I enjoyed Rusty's explanation of the VQGAN + CLIP Unreal Engine trick in Today in Tabs yesterday (subscribers only, sorry). The gist is that people have found a way to improve images created with machine learning by including the phrase "unreal engine" (yes, the Mandalorian-powering video game technology) with their prompt when they ask a GAN to generate an image of that prompt.

Rusty included a link to a colab notebook that could run this particular GAN—VQGAN+CLIP—so I started playing around with it. I know next to nothing about AI or Machine Learning, but I can run code! I tried out the unreal engine trick and the results reminded me of Yves Tanguy paintings.

My next thought was, "why not feed it some lines from surrealist poetry"? (Or their English translations anyway.) So here are a few of those:

VQGAN+CLIP generated image using a surrealist poem phrase, vague castle in a blurry landscape "a meaningless castle rolled along the surface of the earth."
– André Breton, Soluble Fish

VQGAN+CLIP generated image using a surrealist poem phrase, vague night scene with various lights "it is the star struck under my heel in the night."
– Robert Desnos, The Landscape

"The river I have under my tongue"
– Paul Eluard, The River

VQGAN+CLIP generated image using a surrealist poem phrase, a vague flower and vague clock; some fire "where we live the flowers of the clocks catch fire"
– Tristan Tzara, The Great Lament of my Obscurity Three

The other day I linked to Nick Cave saying that AI lacks the nerve to create great music. How about visual art? Seeing machines struggle to make what humans make is interesting, but I don't think the results are inspired. Again, I have no idea how to tune these tools to get more interesting results and maybe further tuning or more attempts would create better images. Right now I think we haven't progressed beyond Hayao Miyazaki's thoughts on art by artificial intelligence. (spoiler alert: "I strongly feel that this is an insult to life itself.")
The Atlantic
"Over the past decade, the blockchain has become a refuge for people who need another place to rest their assets. For global tycoons, it’s just an alternative to parking their money in some real estate they would never visit."
Anil Dash (father of the NFT?!) on its origin and the current state of his problem child.
« Older posts