As Mac'B said, pitot systems work well for the most part, are simple and reliable and understood by pilots, maintainers and designers alike. However, the recent spate of pitot system problems have highlighted a weakness in the design of the "system" in that it relies on multiple-instance of the same technology to achieve the necessary redundancy, and the validity of this reliance has been seriously brought into question.
If pitot icing of some description did overwhelm the system on AF447 it will most likely be declared to be a non-survivable event, one in which all redundant systems have failed. The fallacy in this though is that other similar events have proven survivable as long as there's daylight and a workable natural horizon for reference.
As to the technology change, it's not obvious to me what would be the way to go. One thing I do know, however, is that there's the opportunity for improved combinatorial logic using existing sensor fit that I would guess is being hammered out right now as we speak.
Rgds.
24V