Cambrian-S: Towards Spatial Supersensing in Video - Explained Simply | ArXiv Explained