Eh. Clearly Yahtzee wrote this before the latest rumours came up.
Decent haptic feedback means you can probably make a decent attempt at using a touchscreen without looking at it.
Better still, the one demonstration video shown of that kind of technology shows it can be applied to curved surfaces, and areas that aren't a screen.
This means it can be applied (in what I would consider the ideal case) the grips of the controller.
(Hence, the area where your palm, and most of your fingers are usually in contact with the controller.)
Lots of options for 'extra' dynamic buttons, and lots of options for touch feedback.
Also, yes, stereoscopic 3d is a half-assed trick that is unlikely to ever get past being a gimmick.
But holographic 3d, is according to recent rumours, about a year from a practical (if incredibly expensive) commercial implementation.
And once you have a holographic 3d display, rather than a stereoscopic one, most of the limitations that stop 3D doing anything useful suddenly go away.
You can have proper depth. From something floating right in front of your nose, to something well into the screen at a huge virtual distance...
You won't need glasses. (because if you did, the holographic principle wouldn't actually work anyway)
Your eyes will be able to focus correctly on the light patterns. (because they are reconstructions of the natural way light would fall on your retina for a set of objects with a given 3d relationship to the viewer)
Chances are, because the effect of the most probable implementation relies on real-time processing (because the location of the viewer is part of the calculation), all existing 3d content will probably work on such a holographic display without needing explicit modification.
That means even if such displays start out horrendously expensive, they are a viable 'soft' upgrade from other 3d displays, because unlike moving from 2d to 3d, you don't need to alter the content to do it. (well, not necessarily. - There's probably a mathematically more efficient way of encoding 3d for holographic displays than simply using the method stereoscopic 3d uses for saving such content. - but that doesn't mean 3d content for stereoscopic displays couldn't be tweaked to work. Obviously, content explicitly designed for a holographic display can do without artificial restrictions caused by the limitations of stereoscopic 3d.)
The most serious limitations will be that the viewing angles, and allowable locations of objects (especially if they seem to be in front of the screen) means that the screen edges put a bit of a limit on where things are visible.
The (seemingly) most practical and useful implementation still requires a lot of processing power, and high-speed eye tracking to make it viable. (Both of which together limit the number of people that can watch a single 3D screen.)
And, the requirements of constructing displays in a manner somewhat different to current displays (3d or otherwise) - even if the overall implementation shares a lot of parts with more traditional displays - means the costs of the earliest models are probably going to be astronomical.