Cynically, I think that's it. Optimistically, I like to imagine it will finally be useful in CAD systems, heavy info overlays in real time, realtime graphics simplification in all kinds of neat ways that give us control over the way our senses are perceiving the world we're actively engaging in, but that would require decades of programming that I can't imagine 1, has already been shoehorned into this, and 2, could even be possible given the rapid state of decay in private industry. I think you'd need a government to back an epic 25 year programming project, to take on the likely risk that by the end of decades of work, everything they've done is outdated and irrelevant. Any company is too susceptible to the whims of the market to make something as massive as AR/VR fulfill its potential.