The biggest challenge in interactive video production is that AI models do not distinguish objects in the way traditional 3D models do. A potential solution is the fusion of deterministic compute models with AI train models, which will unlock new opportunities in video production. This integration will enable the rendering of individual objects in a scene using compute models, placing them in an unreal engine, and outputting the video. The next phase in AI model development is likely to involve integrating these two types of models to enhance video production capabilities.