In this video from Vox see how advances in AI research have produced a plethora of deep-learning models capable of generating original images from simple text prompts, and DALL-E 2, Midjourney, and Imagen, are explained.