driver-less cars because they already exist
they are called trolleys
#trolleysrule
One accident and nobody is going to board a pilotless plane.
The possible economical win is much bigger on the car side so I think cars, trucks and busses will be first.
It’s not a tech problem but a regulatory, political and eventually a human problem IMO.
Pilot here.
There’s already a huge amount of automation available for airplanes large and small. The current top of the line will allow the airplane to connect every phase of flight except for the takeoff, coming all the way down to landing on the runway. In your average airline flight, probably 80 to 95% of the flight is flown by computer. The pilots are managing the aircraft, talking to ATC, etc. So you could argue that that is already there.If you mean the ability to conduct a trip without an operator, IE little girl jumps in the back of the car and says ‘Tessie take me to school!’ and the car drives her to school, that will absolutely happen in cars before airplanes. The simple reason is edge cases and emergencies. In a car, if something goes wrong, you simply pull over. Or, worst case scenario, just slow down and stop. It’s not great but it’s not terrible. If something goes wrong in an airplane, you need to keep operating the airplane for anywhere between 10 minutes and 4 hours including a landing. A lot of what pilots do in emergencies is figure out exactly how their airplane has been damaged and strategize around that. A lot of that is intuition, the rest is deduction based on understanding of how the airplane works. Since the computer can’t see out the window or feel things like buffets and sound, a computer won’t necessarily be as good at that. So the pilots aren’t going anywhere.
Technology wise, aircraft are already 90+% automated - autopilot does basically the whole cruise phase, pilots are there to do the communication with ATC, manage the autopilot, and be hands on for taxi, takeoff and landing.
From a legal/policy perspective, the aviation industry is held to a much higher standard of reliability and safety than the automotive industry - the AI driven YOLO that companies like Waymo get away with. It’s not just that autopilot systems have to always work, it’s that they have to always behave in a predictable way.
Airplanes. You can account for natural phenomenon but you can’t account for the guy who decided to randomly swerve to the left cause they felt like it
By tradition, the automobile industry is allowed to build machines that kill people.
The airplane industry isn’t.
This is such a perfect reply lol
Am I crazy? There are already self-driving cars, aren’t there?
No, there really aren’t yet. Driverless taxis and delivery vehicles are all “monitored” remotely by people who effectively drive them when they get into situations the automation can’t handle. Individual self-driving cars all come with a lot of warnings (which many drivers ignore) that they require an active and aware driver for similar reasons.
And Tesla, who have been lying about their self-driving capabilities from day one, continue to run people down and smash into other vehicles on a regular basis.
The systems are good enough to handle 99% of the driving situations they encounter. That remaining 1% is still a long way from being solved. And “pretty good” is not acceptable when failures kill people.
They not working in all cases is a qualifier you are adding yourself though. There are definitely existing self-driving cars. There are no self-driving cars that can handle all situations, but being perfect or finished is not a prerequisite for something existing.
I understand your point, but I disagree. There are currently no cars that are considered fully self-driving as defined by the people who created them. Except for the ones that are really just remotely driven, they all come with warnings that a human the driver must be at the controls and paying attention.
Current self-driving cars are like a printer that works most of the time, but requires a human to read everything it produces and to occasionally write in a few things that it missed or got wrong.
Current self-driving cars
So you agree they exist. You are just saying they are not good. Just like the printer that only works sometimes is still a printer that exists, it’s just bad at being one.
But we are just arguing semantics.
It is mostly semantics. I answered the way I did primarily because I was responding to “There are already self-driving cars, aren’t there?”. That seemed to be asking about functionality, not naming conventions.
I think truck driving is probably the next thing. There’s laws (at least in the US) about how long a driver can run without rest, long haul routes are generally not very crowded with traffic nor complicated. If you can get twice as many hours out of a robot than a human, you can recoup the investment pretty quickly. I could see a hub-and-spoke model where robots handle the long spots with humans taking the busier spokes.
I’m surprised it’s not already in place for rail freight. Pre-defined, well known routes, automatic right-of-way. You’d need some exception detection - spot things on the line or if any part of the train is behaving abnormally, but like cars you can “fail safe” - do an emergency stop if the computer or a remote operator decides that something has gone sufficiently wrong which you can’t do in a plane
It already is for some specific rail freight, iron ore haulage in Western Australia being one example. Rio Tinto has been running them in WA since 2019.
The Sydney Metro is also driverless, albeit a passenger only line rather than freight.
