I think the major point from both accidents is being lost in the technical argument.
The major point is "Fly the Aircraft" Its very nice to know the various control law logic of the Airbus FBW system, just as it is with modern Boeings. However, when the Sh*t hits the fan, fly the plane. If you sit with full back stick in, you're in trouble! In any aircraft, not just modern ones. If you don't know what the biggest control surface on the aircraft is doing then, again, you're in trouble! All this is very easy to say sat behind a computer on a bright sunny day however.
As for fully automated aircraft, show me a computer that can think on its feet in rapidly changing and dynamic situations. A computer that can control multiple failures, passenger problems, ATC issues, weather, failures of unmonitored systems (e.g. tailpipe fire, galley fires). When you can show me that, I will gladly hand in my licence. I will also never step foot on an aircraft again.
Is automation dependancy an issue? Yes, it most certainly is. However, more training is required, not to get rid of the pilots all together.
I will leave you with one final thought. Could we ever design a computer to match the performance of Sullenberger and his crew? I don't mean the landing itself, but could we design a computer that would weigh up all the odds and then make that decision to land in the Hudson? A decision the undoubtedly saved the lives of all 155 people on board that aircraft. My answer would be probably, one day. But not in my lifetime, and probably not for many lifetimes to come.