Thomas Claburn / The Register ·
Netflix debuts VOID, a vision language model that can erase objects from a scene and simulate how remaining objects would behave in the scene without them
Video-language model revises how objects interact when things get removed from a scene