To your last point: we already have such engines, and Google indeed does not do well in those areas. For example, academic research is an area where Google's ta

To your last point: we already have such engines, and Google indeed does not do well in those areas. For example, academic research is an area where Google's take on search does not work as well. But it will take more than the existence of such areas for Google to falter. It would take people in general not being able to find what they are looking for by typing a text query in a box, or finding such a means less efficient than some other. I doubt we will see anything like that before either neural interfaces or devices that can anticipate what you want without any interaction at all.