Generative AI art created with a GAN (Generative Adversarial Network) trained on thousands of artworks that were also created by AI (with CLIP & VQGan). A unique convergence of the two most powerful methods of creating AI art today.

The animation is a traversal through the GAN's 'latent space' - a 512 dimensional computer reality, where every spacial location contains the mathematical seed of a new, original AI artwork.

What you are watching is the movement between successive random 'seeds', creating visions of artworks as we move through the infinite ether of latent space.

As part of an endless series, exploring many different themes and styles, each 'Surreal Vision' will be a randomised and unique collectable, born out of the realm of machine learning and infinite generation.