I think the most interesting part about niantic is how they’ve made hundreds of millions play Pokémon go without niantic ever realizing that the main attraction, wasn’t the AR.

To quote CharlesW's comment below,

>If you think beyond video, Pokémon's AR was a home run.

>My kids know where Pokémon are most likely to live. They know where the gyms are. Their reality has definitely been augmented.

Speaking of the laws of physics, i suspect the primary limitation for this will be the power supply.

I seem to recall that there are ongoing jokes regarding existing AR games that one basically have to bring a generator (aka car) to keep up with the battery drain of the game.

There's a large number of people who will happily walk around with a power bank if there's a good enough reason. This happened for both Pokemon and for Ingress.

I would expect the rendering efficiency to get better fast. The amount of power spent on graphics relative to the actual amount of new information being generated is absurdly high.

We should start seeing more experiences designed specifically for low power renderability soon.

Why do you expect it get better at a significantly higher rate than it has been over the past several decades of computer graphics? Is there some major up-and-coming research on the topic?

No, just a kind of Hegelian turn. Computer graphics have been oriented towards emulating photography for many years. That project is nearing its end and still realism is lacking.

As people start to realize there is more to seeing than just photography, we will enter an era of discovering more sparing uses of computation.

Look to painting for a possible progression. Realism/classicism was just an early maximum. Impressionism followed. And then many other movements. Contemporary painters can do much more with much less.

AR Rocket League!