Hmm. Yeah… well…
I have been hearing about this trend for ages, now. The company I worked in for ten years apparently hated its pilots, and never missed a bender to underline the great corporate motto “no one is indispensable”, be that at an individual level due to your own particular cross section of skills, or extended to an essential group of employees. We were treated to a smug bit of information once, not too long ago, showing us this (not exactly the same video, but the same subject)…
The greater part of the ifs and buts have already been touched on in the posts above, so no need for redundancy. However, in this age when even what used to be the somewhat secluded “plane spotter aficionado brigade” now have a very clear and sound understanding of what it entails to operate and navigate an aircraft, thanks to our excellent simulations, it is plainly obvious even to people who have never been on a flight deck that this simply does not cut it, yet. Sure, FANS and ADS-B are being shunted along to this fully automated end. Sure, it is a prototype, we cannot expect perfection yet (so stop beleaguering us with it, please, until you have something that works in all situations).
And the argument is old, too; thinking, cognizant, trained human pilots are there for the approximately one percent of moments when “routine” goes pear shaped, and some analysis is required to extricate the aircraft from a tight corner. In the usual progress of development (ie; where 99% of problems takes 1% of the effort, and the last 1% of the problems takes 99% of the effort), the “devs” are currently stuck. They might say they are not; geek pride is kind of legendary.
But they are.
Take the A320, for example. All going according to plan, with a flight that ends with a CAT II autoland (even if it is just to keep recent experience), you can switch on the autopilot 5 seconds after unstick and not touch a conventional control again (barring retarding the thrust levers at 20 feet and selecting reversers) until you are stopped on the runway at your destination.
But there are these things called OEBs. A number of them address bugs (yes, straightforward software bugs) that can do all sorts of silly things like throttle back to idle in a climb or revert the whole FCU to basic modes just because the two RAs had a minor disagreement on approach. No one took the trouble to rectify them in the software because, yeah, exactly… 99% effort. Just hand the conditions to the Live-ware, and be done.
Now who do we hand it to? Our friend in the video above, when one of them occupies both seats? They will have their own set of identical bugs between themselves in their DECIDE loop, because they are the same model of android (point for further discussion there, perhaps). CRM is a tool that proactively and deliberately employs the distinct and differing “bugs” of humans to an advantage to solve problems as a team. That will not happen when robots implicitly agree on exactly the same mistake (bug). I know geeks who will laugh and say “this pilot knows nothing” when you say “wait, scrub what I just said, it is wrong”, little knowing that this is exactly what is good about human pilots.
Make no mistake, automation is a great tool. And sure, the day will come when SMS determines all this to be a 1E risk, but, especially to @smokinhole’s apprehension, I wager we will both be long gone, and probably the generation you mention will be retiring, before the see-saw ride ahead this promises becomes a fully fledged reality.
In the meantime, they will keep trying to frighten us. Be interested, and have a chuckle, for what it is worth. At least that is the way I see it.