Earlier quoted context omitted.
It’s my understanding that eye tracking isn’t great as an input method, it should be used more for stuff like rendering or NPC interactions.
I think that eye tracking perhaps could be used to enhance gesture-based input methods though. It could provide a hint at which object a gesture is directed at in cases where that would be ambiguous. I have tried eye tracking as primary input method some years ago in another setting, and I very much did not like the experience.
That's how it works on the Vision Pro. I didn't really feel annoyed by it, but make sure the device is calibrated to your eyes, or the results might not be that great.