A 960M-parameter model turns a still image into a playable world
A researcher has demonstrated a roughly 960-million-parameter model that generates an interactive video world from a single image, taking keyboard controls and live text prompts as it runs. The promising part is that the prompt can change m
Open discussion →