WorldCompass teaches video world models to follow actions better and keep pictures pretty by using reinforcement learning after pretraining.
Spatia is a video generator that keeps a live 3D map of the scene (a point cloud) as its memory while making videos.