Hacker News Reader

Comment by 6stringmerc | Hacker News Reader

> “So many of those features are going into the vehicles today — self-parking, lane keeping, all these things that ultimately are going to be very much a part of full autonomy. And customers are already getting used to it.”

This reminds me a lot of how advanced and amazing the avionics and support systems are on Airbus planes, and why shit pilots crash them whenever anything doesn't go exactly like a simulator textbook. It's a part of the future of automotive "advancement" that I'm not optimistic about. This is certainly a personal opinion, one coming from a guy who likes to really be involved while driving and has his fair share of mistakes on his record, but I think it's worthwhile to keep in mind.

View on HN

> and why shit pilots crash them whenever anything doesn't go exactly like a simulator textbook.

Often it according to a simulator textbook but pilots haven't studied the text book enough so they end up being confused what mode the autopilot is on, override and crash the plane.

I think FAA even calls that as such -- "mode confusion".

Of course it could be the problem of user interface or warning and error reporting is not clear and explicit (alarm is blaring so they ... just ignore it because it has so many false positives).

Don Norman would claim it is the designer who is at fault:

http://www.amazon.com/Design-Everyday-Things-Revised-Expande...

Interesting discussion in itself.

Upon further reflection, I think some shit pilots still manage to crash the plane even when something IS a textbook simulator issue.

That shorting of the runway in San Fran? Every warning worked as intended except the pilot. The prop plane that went down in Taiwan? Pilot shut down the wrong engine.

Basically if we're talking about a person being unable to actually handle the needs of modern piloting - mode confusion as it were - then I'm suspicious that they aren't very good pilots in the first place. This is where I see the correlative in cars.

Just because lane-alert and auto cruise control can help an inattentive driver stay out of a crash doesn't mean the systems can't be compromised by a poor operator. That's kind of my thought process.

There's a really good podcast over at 99% invisible called Children of the Magenta that deals with this. IIRC, he (Roman Mars) also did a follow-on show about what it means for driverless cars.

And the number of accidents with advanced systems + shit pilots is still lower than that of good pilots + no advanced systems.

That's also worth keeping in mind.

What about advanced systems - pilots?

Nobody is arguing against radar altimeter, weather radar, gyrocompass and artificial horizon.

I'm very doubtful that the actual autopilot has saved any lives. It's more likely to just allow captain to nap.

Tired pilots are very dangerous. So, the actual numbers on this might be counter intuitive. We may very well be at a point where flights would be safer without on board pilots.

I'd like a citation of a study or some numbers that back up your claim, really I would. I'm strictly speaking about commercial aviation - not puttering around in a Cirrus with a parachute. Those aren't equitable comparisons.

Mostly because in recent times, pilot error has been the most frequent cause of crashes, to my knowledge, in Airbus planes. Whether that's being too stupid to avoid an active combat zone and get shot down, or turning off the wrong engine and causing a prop plane to crash in a recoverable situation (granted not an Airbus).

Basically, there are pilots flying in commercial aviation today that, practically speaking, are not qualified to be stick & yoke pilots but are commercial pilots because demand far outstrips supply in some parts of the world.

It wasn't too long ago in this country that planes were crashing all the damn time. Go here: https://en.wikipedia.org/wiki/List_of_aircraft_accidents_and... then sort by date and you can see the number of incidents in the US just evaporate.

Are you seriously suggesting Airbus aircraft are more dangerous than Boeing due to automation? The (public) accident data show that they have comparable safety records across the board. Carriers aren't clamoring to Airbus for less automation. Flyers aren't gripped by fear when they see "A320" on their boarding passes. The manufacturers employ different philosophies on flight control, but it's not a clear shot that a human is always better. With cars, the general data suggest that automation leads to massive reduction in accidents. In every field there's some nostalgia and romanticizing over how when SHTF, only those with manual experience will be able to achieve anything. Then you have cases like the death of pilot of Michael Alsbury, which would have been prevented by very simple automated flight control checks. http://www.latimes.com/business/la-fi-ntsb-virgin-20150727-s...

No, I'm seriously suggesting that some airlines prefer to purchase Airbus products because their advanced avionics and automation systems bridge a competency gap in order to maintain operations and growth at the expense of safety.

Are you seriously suggesting a traveler of reasonable means will look at their ticket that says AirAsia on it and think they have made the best possible deicison?

Reply on news.ycombinator.com