The first one is mainly solved, the second one is way harder due to head tracking latency, computer vision latency, and user-dependent eye projection.