Imagining that the slow glass is really just a huge chunk of flash memory with imaging devices on both sides…
10 years of glass is described as being 1/4 inch thick. Let’s suppose that we want resolution equivalent to a large HDTV, so I’ll assume pixels 0.5 mm on a side. That gives us 1.5 mm[sup]3[/sup] of volume for each pixel.
Since it’s holographic, each pixel actually stores an image that is projected out to its hemisphere view. Since we aren’t going for perfect resolution, I’ll just assume it needs 1000x1000 pixels (at 60 fps).
Modern flash memory has a cell size of perhaps 40x40x40 nm. It’s already stackable, though not quite to the extent of supporting giant slabs of the stuff, but I’ll assume we can get to that point. If each cell can hold 3 bits, our 1.5 mm[sup]3[/sup] pixel chunk should store around 7e13 bits of data.
10 years of video then needs 1.9e10 frames, or 4.5e17 bits uncompressed. So we need 6400x compression, which sounds like a lot, but really isn’t so bad–h.265 does a reasonable job at 2000x compression. And this isn’t taking into account the benefits from adjacent cells. Not to mention that unlike a movie, the average scene doesn’t have much going on most of the time. Nearly static scenes compress very well.
So the technology isn’t totally nuts, at least in a somewhat primitive form. And I’m making pretty conservative assumptions here; there’s no physical reason why storage can’t increase by a factor of 1000x, or even (optimistically) by 1000000x.