A multimodal AI that can generate novel videos with text, images or video clips | Hacker News Reader