At the risk of grossly oversimplifying a complicated situation, it would seem we need to get away from the current "train to fly the instruments" and get back to "train to fly the aircraft"
Automation is without doubt our friend, but apparently only to a point.
We must perform a risk assessment, do the benefits of automation outweigh the risks? I think that answer is a clear yes - the balance of positives far outweighs the few negative results as here.
Normally a rational risk assessment would say Asiana 214 is an anomaly, certainly necessary to review and consider, but one incident should not outweigh years of safe ops.
But the Asiana failure was so massive and complete - the complete failure to simply fly the aircraft in CAVU conditions and with a full flight deck of allegedly experienced pilots - that the level of the failure was so basic it shocks our belief. And escalates the importance in any risk assessment exponentially.
It seems a conundrum.
Do we require more hand flying, without automatic aids, which may slightly reduce safety in the short term, but likely increase safety in the long term. Do we require a bunch of hand flying time in the sims to try and train out the automation reliance? Do we take away some of the automation altogether - reduce the complexity - in the interests of safety?
Do we FURTHER automate, attempt to "tech" away the problem? Do we require "certification" to fly into major and/or "difficult" airports? Do we require all pilots be trained and certified in a unrelated stand alone program outside the company? Do we require one pilot to wear a hat that says "I'm in charge" so there is no uncertainty?
Some of these are obviously silly. But they reflect the complexity and importance of the issue.
The failure was at the most basic simple level. The solution seems anything but.
It seems to me there absolutely IS a cultural issue. Crew hierarchy and the tendency to train (and/or learn) by rote repetition and memorization. And this must be addressed.
But the automation "trap" seems to transcend that - we are beginning to see indications this fault could affect all of the system. And that would seem to be the bigger elephant in the room.
To me a couple short term changes might be a help in at least identifying weak competency at hand flying, and removing company/culture bias.
What if certification was required to be completed by independent authorized arms length organizations? And what if every certification required a hand flown, no instrument/automation, approach and landing, preferably in the aircraft?