With some AI/NERF magic, you could adjust IPD and allow limited head motion for videos and photos, even when they're shot with a sub-optimal stereo setup or even on a single camera.
sure, there are already models that allow you to hallucinate a full 3D scene from a single image. Still, it would be pretty cheap to add that camera I assume and it would have the potential to pretty quickly fix the 3D content issue that is prevalent currently without introducing hallucination artifacts.
