When digital artist Robbie Barrat handed out free NFT coupons at Christie's four years ago, most guests dumped them in the bin, not realising they would soon be worth millions of dollars.
Barrat, then still in his teens, had been invited by the London auction house to talk about the rise of online art.
As part of the presentation, he gifted the crowd 300 cards, each with a code that gave them rights to a digital artwork he had created using artificial intelligence.
This was before the NFT market exploded last year, and so only about two dozen of the guests bothered holding on to their little cards.
Barrat later recovered many from garbage cans and the floor.
On March 2 this year, just one of those artworks, "Nude Portrait#7Frame#64", was sold at Sotheby's for £630,000 ($821,000).
Barrat, now 22, had been working with AI since high school in the United States.
He made his images by uploading 10,000 nude images from classical art into his computer and then using two competing AI programmes to distort them.
"My interest was: can I use this tool to make something that is not classical?" he told AFP in a video interview.
The method is known as "generative adversarial networks" (GANs): two neural networks that compete with each other using algorithms.
"(They) sort of fight between each other," Barrat said, adding that he purposefully added glitches to the programmes to make the final results more interesting.
The result was a series of shapeless "nudes", unsettling masses of reddish and brown tones that bear resemblance to paintings by Salvador Dali or Francis Bacon.
Barrat was invited to talk at Christie's by art collector Jason Bailey, known in the crypto-art world as Artnome, one of the pioneers of the NFT market.
"Nobody knew what an
Read more on tech.hindustantimes.com