ELAC,
I'm not trying to downplay the benefits of automation in any way, just the total apparent reliance that is becoming more and more evident. As automation takes on more and more of the work load it's readily apparent that the system is both to complex for at least a portion of the user base to manage under stress and that it's use is compromising basic skills to some significant measure.
Instead of automation leading to a net increase what we're seeing is an increase in the capability of one segment (automation) leading to the degradation of another (actual hand flying capabilities).
We're at a point where its possible for a pilot to log 80 flight hours and have less then 2 or 3 total hours (if that) actually hand flying a plane. Where I disagree with you in your interpretation of "decision making". That is exactly what hand flying is, making decisions. The alternative is observing a computer automatically make those minute corrections for you. After all isn't that what software does? It takes the raw data stream and applies a set of predetermined formula's and acts based on the resulting calculations.
If in fact AF447 was an upset at altitude due to the automation reaching a point it kicked out (pure speculation) then would not the lack of hand flying experience at altitude have severely hampered any attempts at recovery by the pilot flying since he had no real experience under such conditions.
This further compounded by a confidence in the automation that encouraged an altitude selection beyond what a hand flying pilot would be comfortable maintaining given the very narrow boundaries?
At the end of the day how can a competent pilot rely on automation "to little" . Nothing can be safer (and yes thats just my opinion) then actually hand flying an airplane. Otherwise lets just equip em all with a little "autosave" box...
Automation in place of training and proficiency is going to cause significant issues...just one persons opinion...
In event of emergency break glass, pull lever and sit back and enjoy the ride