>* If the screen, camera, and speed were good enough in 2016*
You say it's a rhetorical question but I'm not clear on why or why this sentiment is so persistent. After all, the screen, camera, and speed absolutely WEREN'T good enough in 2016, any more than regular computers were good enough in 1986, 1996, 2006 or 2016. They were simply what could be managed at the time with technology at the time. The only aspect of electronics that is "done" for typical audiences [0] is audio, where we have microphones, recording and reproduction that can (easily) exceed the biological limits of human hearing. In contrast exceeding human visual acuity in capture, storage and reproduction remains a work in progress (though it's conceivable we'll hit it in the next decade or so which will be a very interesting change for our industry). That in turn itself drives some demand for computation, storage and processing, though more fundamentally it's hard to say if there is any real limit on how much computation might be put to use. Storage has been on a fast enough upward curve that I think it might be said it's approaching the point where regular people always have enough merely in the course of normal upgrades, but to handle an entire lifetime.
So yeah, come back in 2032 maybe.
----
0: Scientific applications of course are frequently interested in sounds that well exceed human limits, though even there we have the tech for it albeit not in non-specialized devices.