Text prompts can’t deliver real VR control. This explains how hand+head pose conditioning makes generated worlds respond to your actual movements.