It is pieces of the puzzle really. You need machine learning, computer vision, robotics and ultimately AI plus some kind of self-assembling machines to really make it work.

We don't have the tech yet. But it is getting there. Being able to find the translation of a spoken phrase, or answer a spoken question accurately is actually difficult. You pretty much need your phone to be aware of the environment; if you want to get good results that is. So projects like Siri, Google Now or Glass are actually important and potentially can be crucial and 'world changing'.