"The account has produced everything from
a version of Salvador Dalí’s painting “The Persistence of Memory” in the neon-pastel style of Lisa Frank to
a depiction of 'the edge of reality and time,' a frightening swirl of floating eyes, hourglasses, and windows onto nowhere.... [T]he account’s proprietor is... Sam Burton-King, a twenty-year-old student at Northwestern University.... They began as a math major, but, finding the coursework too difficult, moved into studying philosophy and music.... To create art for @images_ai, Burton-King feeds a selection of written prompts into what’s known as a generative adversarial network (G.A.N.), a machine-learning system in which two artificial neural networks—computer models that mimic the information processing of a human brain—compete with each other to come up with a result that best matches the inquiry..... 'People ask for Shrek all the time, or
Big Chungus, or Donald Trump in various situations,' they said. One successful prompt was
'Elon Musk experiencing pain'.... Burton-King feels that the looming impact of A.I. is out of their hands, though; the machines are getting more skillful all the time.....
'I don’t mind contributing to the artificial-intelligence apocalypse,' they said, 'as long as they have a cool civilization after ours.'"
From "Appreciating the Poetic Misunderstandings of A.I. Art/The Twitter account @images_ai has gained a following for its feed of surreal, glitchy, sometimes beautiful images created through machine learning" by Kyle Chayka (The New Yorker).
Posts
6 comments:
Well, I hit a speed bump at "They began as a math major,..." I went back to see who's name I missed being paired up with Burton-King. But, no. Sam Burton-King is not sure what or who or how many he is. So he's probably the right person to be working on extruding a new reality through computers. He's using a generative adversarial network which, as they state it 'mimics the information processing of a human brain'.
Not surprisingly, this version of the human brain wants to see Elon Musk experiencing pain, or Donald Trump in various hideous contortions, among others. It's all interesting, but misses how artists create art. Still, that's never stopped anyone from producing faux art, kitsch art, or knock-offs at various points of humanity (remember velvet Elvis?).
All that said, this is very interesting and will have an impact on future art. Perhaps we'll find that there is a small, niche market for real art from real humans, using only a brush, a canvas, some colors, and an actual human imagination. Really, we're currently using our imagination to create a synthetic imagination that will soon surpass our own information processing capabilities and remove us from the equation.
At that point neither they or me will matter.
That "art deco buddhist temple" looks like a bouncy house at a shopping mall in Dubai.
Looks it was inspired by the Djose Temple in the Final Fantasy X video game. Has that vibe totally.
Math is hard. Feeding prompts into a machine you don't understand is easy. Spin art for the current decade.
This is the second post on AI-generated "art" in a month - or maybe it's the first Althouse post to set aside preconceptions, and take this human endeavour seriously.
The first was on July 15 - and I still think this particular 'AI art' is visually quite pleasing - despite being given the Althouse label of "bad art".
It looks like a set from Big Trouble in Little China. Where is Jack Burton when you need him?
Post a Comment