Video made by text-to-image machine learning methods, based on CLIP & VQGAN. Text input = "Cubist self-portrait". Special thanks to @advadnoun for Latent3Visions! ✌️