We also have a happy habit of over-embracing the new. Radium cures, lead in petrol, Heroin, Thalidomide, electropathy (being buzzed by high voltages for the "health" benefits), arsenic pills to increase libido.
It's only some years later we realise whether it was a great idea, a really stupid one or a straight up con. So a little reluctance and wariness, especially in a world where everything is marketed as being a brilliant idea, is probably a very good thing.
Self driving cars are definitely in the don't know yet category for instance.
The problem with self driving cars is they aren't. Too many asterisks and fine prints. You cannot jump across pit in two jumps. And that is exactly what they are trying to do now.
The car is self-driving until the moment it isn't. And being aware of your surroundings but not driving is extremely exhausting. And you have just a second or two to react.
The case history of highly trained vehicle operators, assigned at several per vehicle, failing to respond appropriately, over a course of several minutes to a failure of self-driving mechanisms is particularly sobering.
Yup, that air france flight and I would guess many train incidents too.
I'm surprised we do not compare self driving cars to trains more often. Trains are on tracks and still require drivers who have to hold onto deadman switches.
Full automation of cars with no driver are not coming any time soon.
3. Exceptionally high degrees of training. These are skilled professional pilots, specifically drilled in mishaps, and independently aware and capable of discussing or diagnosing the situation.
Multiple modes of failure occurred (it's a fascinating case study), starting from bad sensor design (prone to icing) and weather conditions (icing) leading to loss of airspeed indication. Lack of feedback between pilot and copilot controls, and input-averaging, combined with failure of the co-pilot to relinquish control of aircraft to the pilot, led directly to control-and-response failures which ended in impact of aircraft with the sea at over 100 knots each of vertical and horizontal velocity components.
What it wasn't was some distracted inexperienced teenager crusing in daddy's Autopilot Tesla whilst distracted on mobile calls, texts, and showing off to other passengers.
Trains and tracked vehicles are another good case in point.
Automobile autopilots will have to respond to degraded conditions by adapting to safer modes -- lower speeds and recognising that they're operating out of their design parameters. It doesn't seem Tesla are doing this. Google's approach seems more conservative.
> Automobile autopilots will have to respond to degraded conditions by adapting to safer modes -- lower speeds and recognising that they're operating out of their design parameters. It doesn't seem Tesla are doing this. Google's approach seems more conservative.
Yeah. Tesla seems pretty committed to enabling hands off driving for minutes at a time.
The self driving, especially Tesla, is being promoted as being along the lines of "do whatever you like, the car will cope". The public, rightly or wrongly, are forming the impression they can read a book, chat with passengers, do makeup, eat lunch and completely ignore what's happening outside.
What it actually is, of course, is cruise control plus. If it were promoted as such we'd be less likely to have folks making hands off videos trying to find the limits of the system.
When, as will undoubtedly happen, we get true self-driving we're going to have to accept new failure modes and deaths as a result. People will die as a result of AIs going out of known parameters. Perhaps innocent pedestrians, perhaps we'll find they're especially bad with children or cyclists. We may require legislation that in certain weather, road or other conditions self-driving must be disabled.
It took an awful lot of years, miles and deaths to get to the exceptionally safety conscious aviation we now have. If we're just passengers will we start to expect a similar level of safety with self-driving?
Sitting in the back of the magic car with iPad whilst it takes you from any door to any door is, sensibly, still a good way away.
It's only some years later we realise whether it was a great idea, a really stupid one or a straight up con. So a little reluctance and wariness, especially in a world where everything is marketed as being a brilliant idea, is probably a very good thing.
Self driving cars are definitely in the don't know yet category for instance.
A firmly one-sided article promoting a new book.