This is good news for VR (or spatial computing). If the models understand the physical world as well as the paper shows, generating two projections of a scene d

This is good news for VR (or spatial computing). If the models understand the physical world as well as the paper shows, generating two projections of a scene does not sound like a difficult ask. Really excited for what's to come.