The interface world model can also generate audio, allowing for sound effects tied to user actions