3 ms·

I find myself in the strange situation where I both agree and disagree with this sentiment. The source of disagreement can be described as "time" - I consider

I find myself in the strange situation where I both agree and disagree with this sentiment. The source of disagreement can be described as "time" - I consider not only the past, but the delay that happens after a technology is described to when it makes changes in our lives. It all takes time, and in the time to come, there is a chance that autopilots can take more responsibility from the pilot. Would you really be confident that in 20 or 30 years time, pilots will have the same job description as of today? Of course, a pilot in a passenger plane has a lot more responsibility than simply piloting the plane. For example, dealing with passengers or talking to ground control. The autonomous trucks will not take driver's seat overnight - the truck in question will be a decade-long testing phase, according to the article. What happens afterwards is anyone's guess.

ubernostrum 11y ago
Would you really be confident that in 20 or 30 years time, pilots will have the same job description as of today? Yes. "Autopilot" systems are, in terms of capabilities, well behind even current widely-deployed technology in ground vehicles. In the air, an autopilot system is essentially glorified cruise control, and nobody asks when cruise control in cars will take over for humans. Autoland systems are pretty far behind automatic parallel-parking features in modern cars; autoland requires significant human decision-making, human input and human oversight to function. And the purpose of autopilot/autoland is not to replace the pilot, but to relieve the pilot of mundane tasks the software can do perfectly well, freeing the pilot to do things software still is -- and likely will continue to be -- bad at. Unless, of course, you think friendly human-serving general-purpose AI will be available within the next two or three decades (I don't).
- greendesk 11y ago
  >>Unless, of course, you think friendly human-serving general-purpose AI will be available within the next two or three decades (I don't). /sigh. A decade ago I started my bachelors. When I was in college, I expected that general-purpose AI will be readily available by this point. How naive I was! Currently, I doubt it will be available by that point in time. /sigh I do not know how much progress will be achieved within the next two or three decades. In the circumstances of passenger flights, keeping current solutions is likely a conservative and more pragmatic approach. With regards to my comment, I do not expect autopilots to substitute pilots. Instead, I expect that they should be able to handle additional tasks assigned from the pilots. For example, it was not a passenger flight, but a helicopter crashed in Nepal during the rescue from the earthquake. I wonder if its pilot had fore-warning from sensory information prior to the crash; and if they had not, how to design its warning system and if an autopilot could take control of the helicopter to land it safely. The sources I have read on passenger pilots use of autopilots is: http://flyingforeveryone.blogspot.com/2012/01/autopilot-myth-what-your-pilot-really.html http://flyingforeveryone.blogspot.com/2012/01/autopilot-myth... http://aviation.stackexchange.com/questions/1802/why-do-we-still-use-pilots-to-fly-airplanes http://aviation.stackexchange.com/questions/1802/why-do-we-s...

ubernostrum 11y ago

Would you really be confident that in 20 or 30 years time, pilots will have the same job description as of today? Yes. "Autopilot" systems are, in terms of capabilities, well behind even current widely-deployed technology in ground vehicles. In the air, an autopilot system is essentially glorified cruise control, and nobody asks when cruise control in cars will take over for humans. Autoland systems are pretty far behind automatic parallel-parking features in modern cars; autoland requires significant human decision-making, human input and human oversight to function. And the purpose of autopilot/autoland is not to replace the pilot, but to relieve the pilot of mundane tasks the software can do perfectly well, freeing the pilot to do things software still is -- and likely will continue to be -- bad at. Unless, of course, you think friendly human-serving general-purpose AI will be available within the next two or three decades (I don't).

greendesk 11y ago
>>Unless, of course, you think friendly human-serving general-purpose AI will be available within the next two or three decades (I don't). /sigh. A decade ago I started my bachelors. When I was in college, I expected that general-purpose AI will be readily available by this point. How naive I was! Currently, I doubt it will be available by that point in time. /sigh I do not know how much progress will be achieved within the next two or three decades. In the circumstances of passenger flights, keeping current solutions is likely a conservative and more pragmatic approach. With regards to my comment, I do not expect autopilots to substitute pilots. Instead, I expect that they should be able to handle additional tasks assigned from the pilots. For example, it was not a passenger flight, but a helicopter crashed in Nepal during the rescue from the earthquake. I wonder if its pilot had fore-warning from sensory information prior to the crash; and if they had not, how to design its warning system and if an autopilot could take control of the helicopter to land it safely. The sources I have read on passenger pilots use of autopilots is: http://flyingforeveryone.blogspot.com/2012/01/autopilot-myth-what-your-pilot-really.html http://flyingforeveryone.blogspot.com/2012/01/autopilot-myth... http://aviation.stackexchange.com/questions/1802/why-do-we-still-use-pilots-to-fly-airplanes http://aviation.stackexchange.com/questions/1802/why-do-we-s...