There's also less to render, if you only render where you're looking. You can render lod proportional to retina receptor density. Can be done with eye-tracking

There's also less to render, if you only render where you're looking. You can render lod proportional to retina receptor density. Can be done with eye-tracking, or contact lenses. Big the big problem is latency: it has to be even quicker than what current VR headsets can get away with, especially for saccades. We may have direct visual neural implants before we have that.