Not everyone working in Machine Learning thinks General AI is even theoretically possible, let alone imminent. Like myself, some think machines will never win the Imitation Game (aka The Turing Test).
Anyway, the point being that if humans to not master the easy and medium parts of a system, they can never understand the difficult parts.
An example would be the stupid idea that a human can sit in a car and take over when the alarm sounds because the car doesn't understand what's happening. Unless the human was paying total attention then it is a hopeless situation, he does not have a feel for the road conditions, was not observing the environment, is not "warmed up" to operate the controls and needs to do all those things because the car is about to crash!
I have experienced a similar sensation myself, turning the cruise control off as I leave the motorway I always take a moment to reacquaint myself with the position of the pedals (we have 3 of course, with a clutch, not automatics) and give the brakes a squeeze before I need them at the intersection I am speeding towards.
Same goes in a factory control environment, unless you have solved the easy problems a sudden requirement to avoid disaster will be hopeless.
But this was all anticipated by E.M. Forster in 1909 in his short story The Machine Stops which was written as a counter to the hubris of the H.G. Wells type techno utopians.
You might also enjoy The Technological Society by Jaques Ellul (1963) although that is more in-depth. It was Ted Kaczynski's bible (according to his brother) and is a required reading for those who wish to call themselves Neo-Luddites.
All I want is a toilet that wont flush unless the lid is down.
Ellul's premise is that we are going down a path of deskilling where optimisation of the process is the goal and humans have become prisoners of the machines under the weight of consumption, which distracts us from the essence of humanity.
A bit of a black pill for me, as I specialise in process optimisation.
Not everyone working in Machine Learning thinks General AI is even theoretically possible, let alone imminent. Like myself, some think machines will never win the Imitation Game (aka The Turing Test).
The animal brain is not clockwork.
There's an 1983 paper, The Ironies of Automation
Anyway, the point being that if humans to not master the easy and medium parts of a system, they can never understand the difficult parts.
An example would be the stupid idea that a human can sit in a car and take over when the alarm sounds because the car doesn't understand what's happening. Unless the human was paying total attention then it is a hopeless situation, he does not have a feel for the road conditions, was not observing the environment, is not "warmed up" to operate the controls and needs to do all those things because the car is about to crash!
I have experienced a similar sensation myself, turning the cruise control off as I leave the motorway I always take a moment to reacquaint myself with the position of the pedals (we have 3 of course, with a clutch, not automatics) and give the brakes a squeeze before I need them at the intersection I am speeding towards.
Same goes in a factory control environment, unless you have solved the easy problems a sudden requirement to avoid disaster will be hopeless.
But this was all anticipated by E.M. Forster in 1909 in his short story The Machine Stops which was written as a counter to the hubris of the H.G. Wells type techno utopians.
You might also enjoy The Technological Society by Jaques Ellul (1963) although that is more in-depth. It was Ted Kaczynski's bible (according to his brother) and is a required reading for those who wish to call themselves Neo-Luddites.
All I want is a toilet that wont flush unless the lid is down.
Ellul's premise is that we are going down a path of deskilling where optimisation of the process is the goal and humans have become prisoners of the machines under the weight of consumption, which distracts us from the essence of humanity.
A bit of a black pill for me, as I specialise in process optimisation.