My point about AF447 was:

  - even with redundancy, there can be enough failures that the computer no longer has enough information to make the right decision, one thing humans seem to be good at.

  - people who had been trained in stall recognition and recovery, but probably had not kept proficient at it (unlike pilot of slow airplanes like GA and glider pilots) failed to recognize it (only 1 of the 3 pilot did) and did not recover from it.