asimovian visions uses gpt and stable diffusion to picture the everyday future scenes asimov cherished. each artwork is a tile of a parent scene, generated with controlnet.

unique qr codes were deployed at nfc lisbon 2023 so that visitors could mint. more than 200 artworks were minted in less than an hour.

i wanted to translate asimov's vision of the future into the current state of ai. gpt-3.5 was the discussion partner: it generated ideas, descriptions, and titles for the collection. stable diffusion 1.5 turned those ideas into images, iteration by iteration.

categories

• * ai in a world of innovation and creativity
• * ai in a world of ethics and identity
• * ai in a world of education and infrastructure

each parent concept expands into a mosaic of specific subjects within its category. with controlnet, every nft emerges from the parent visual, which guides color, structure, and meaning.

the pieces were shown at nfc lisbon 2023 and were at first mintable only by visitors. to mint, you scanned a qr code embedded in the mosaic, which led to the mint page.

the qr codes contain the artwork itself. i built them by hand, with a technique few people know.