AR is a white elephant.
First, the resolution of AR displays will be far lower than that of traditional displays. And that's not gonna change any time soon.
Second, to render AR content in a way that will not cause eye strain (e.g. it renders content that you can focus on like any other object in your field of view), you need a proper light field display. That tech doesn't exist.
Light field displays are certainly an area of study at major institutions, but we aren't on the cusp of suddenly having commercially available light field displays.
Third, ergonomics. You need a powerful computer to render this stuff. That means lots of energy, which means hot. It also means big batteries, which means heavy. You're not going to use AR glasses as a computing platform if you can't wear them for more than an hour at a time.
Fourth, cost. All of this tech is gonna be pricey. Magic Leap has said their glasses are going to be at least $1000. HoloLens has been available for a while now for $3,000. People aren't going to buy *another* $1000 gadget. The only way this would work...
...is if your AR glasses truly replace your traditional computing devices. Which, it can't, because see the other reasons in this thread.
Finally, input. If you can't type an email comfortably with it, it's not a computing platform. It's a gimmick. And I'd love to hear you explain how we can program in an AR environment if we can't even type an email. 🙄
If we have to pull out our smartphone to act as a keyboard, why are we even buying AR glasses in the first place?
This is what a keyboard in AR would look like. It would also function similarly. Did it work back then? 
