I mean I get it, but for me it just doesn't seem to sound as "real" as people claim it does. This appeared on my newsfeed earlier, apparently people are getting excited about binaural audio again, but to me it just doesn't really sound any more spacial or realistic than a good stereo recording. 
 
http://www.theverge.com/2...-times-square-new-york

 
It would be really nice to be able to build a sonic scape in which every sound object was clearly located in three dimensions.  And then be able to have each object built from smaller sound elements so you could, for instance, have one sound at xyz with another sound looping around it.  The problem is we use much more than our ears to understand the sonic world around us (eyes, head position, chest, skin, etc., etc.).
 
I played around a little bit with this last summer and just getting one stereo synth sound (say with some elementary panning going on within the sound) narrowed down to an 'object' that can then itself be panned around is not an easy feat.  The intention was then to move on to up/down and front/back playing with EQ, delays and such.
 
Even if it were possible to create something interesting, chances are it'd only work for your own head as everyone has there own HRTF (Head-Related-Transfer-Function).

Binaural is meant mainly for headphone use.  Its use is rather limited.  Still some people swear by it, like some do with MS.  I just shrug it off and try to make recordings that sound as good as possible using whatever is most appropriate.