Multimodal models and video infrastructure for physical AI
Video understanding models that search and analyse what happens in video
