To me, the real significance of a completely successful attempt by Hollywood to make an all-CG film that doesn’t seem like computer graphics is this:
Whenever we see some level of special effects in a movie, what we are seeing is the cutting edge of what can be computed in about an hour of compute-server time (more or less). Then, about a decade or so later, that level of computation shows up in computer games.
So it takes about ten years for some level of computer graphic realism to go from “that took about an hour” to “that took about 1/60 second”. The transition is not as difficult as you might think, because games, unlike movies, can take advantage of special purpose graphics hardware. Because the requirements of movies to be true to reality are relatively high, they can’t take advantage of the latest in special-purpose hardware acceleration. So compared with game graphics, movie graphics lose up a factor of 100 in efficiency (which they gain back in flexibility and generality of effects).
What all this means, in the larger picture, is that sometime in the next ten years your consumer-grade augmented reality glasses will be able to simulate visual reality more or less as well as the film “Gravity” does now. This means that you will be able to look around and see a transformed reality, if you like, which is indistinguishable from the real thing.
Except, maybe, for faces.
You’re right about the faces. I thought the alien faces in Avatar were pretty good until later seeing a “how they made it” clip on YouTube. The clip showed the computer generated face side-by-side with the actor’s actual performance. No comparison. You can’t un-see it…now the Na’vi look fake to me.