This is insanely cool. Near real time translations subtitled and beamed onto your retina. Who doesn't want one? while I get the impression this will (initially) be like the voice recognition software of the late 90's (ie requires a fair bit of work for 50/50 results) it has several features that are very significant.
The actual translation system is the not too surprising amalgamation of voice recognition and translation software. While both these technologies have been around in vague forms for the last 10 or so years it has only been in the last 5 years that they have become truly viable. Voice recognition interfaces are pretty common now on dial in services (eg for national rail) and while they do get it wrong from time to time they are pretty good. The same goes for translation, most things can be translated by, for example, google translate; it may have some interesting turns of phrase but the fundamental message gets through and this is a free, web-based tool.
The most exciting feature of this though is the retinal projection, this is something that has been slowly emerging of the last year or so (either as micro screens or direct projection) and is one of the last technologies needed to achieve complete augmented reality. We nearly have complete internet data coverage (it may not be fast but there are fewer and fewer places where I can't get some sort of net) and with technologies like 4G, BT openWorld (all BT customers can jump onto free open net connections from BT) or the cloud all poised to become very common it shouldn't be too long before people are wondering how they coped without being able to access wikipedia while walking down a street.
Augmented reality is a while off yet (and still limited by battery life) but as more technologies start to use it (see the phone app layers) it is fast becoming reality. I would say that within the next 10 years at the latest people will have augmented reality headsets (most likely glasses) and that they will be common within 15 years. As a technology I suspect that it will be similar to voice recognition in that for the next year or so everyone will be claiming to use it with no actual utility. After that, and especially with things like the translator specs, it will start to re-emerge as general utility, to steal an idea from Charles Stross: wouldn't it be cool to have a map of the city your in over-laid on your vision and then be able to watch the progress of your bus towards you?
This is, I think, the most significant part of these glasses: not so much what they can do (which is very cool) but that they represent a protrusion of what so far as been a fun toy in labs into useful utility.