Imagine it were 90% automated. Now imagine there's a 3 hour outage of the automated system.
You're left with a bunch of planes in the sky that can't stay there forever, and not enough humans on the ground to manually land them.
Now image the outage is also happening at all airports nearby, preventing planes from diverting.
How do you get the planes out of the sky? Not enough humans to do it manually.
Now imagine the system comes back online. Does it know how to handle a crisis scenario where you have dozens of planes overhead, each about to run out of fuel? Hopefully someone thought of that edge case.
Remember when all the Waymos were confused by a power outage? Now do that, but with airplanes that will fall thousands of feet and kill hundreds instead of park in the middle of the street.
I'm not saying we shouldn't automate things. We should. But, it's not easy. If it was, we would have done it already.
I think the point they're making is that the failure mode of a waymo and automated air traffic control could look the same from an angle, but would have very different consequences.
That's what everyone screaming 'funding' doesn't seem to understand here. If your failure mode for potentially hundreds of people dying is one controller over radio forgetting something, then it'll happen eventually. And has happened, there's plenty of videos on youtube of near miss radio recordings. When a plan is landing at over 100mph simple good luck can take care of things the majority of the time.
It just feels wrong that the primary form of control in 2026 is voice over radio.
> Now imagine there's a 3 hour outage of the automated system.
Planes divert to another airport, passengers grumble, end of story. Airport closures can and do happen all the time for all kinds of reasons, including weather or equipment malfunctions.
Speaking of runway crossings specifically, you could have an automated backup, and require authorization from both ATC and the automated system to enter a runway.
We build pacemakers, AEDs, flight control software, and other mission-critical life-and-death software. The idea that we'll just forever keep the system run by specially trained humans with known and foreseeable faults because poorly designed software could fail is head-in-sand unreasonable.
Look what happened when the power went out in SF and the Waymos just stopped in the street because they were confused and there weren’t enough humans to direct them. Now imagine that but with planes that will fall out of the sky when they run out of fuel since they can’t land. Automating this is pants on head retarded.
That sounds like a poorly thought-out implementation.
An example of a poorly thought-out implementation elsewhere does not exclude the possibility of coming up with a better one than humans coordinating with their mouths over radio.
You're left with a bunch of planes in the sky that can't stay there forever, and not enough humans on the ground to manually land them.
Now image the outage is also happening at all airports nearby, preventing planes from diverting.
How do you get the planes out of the sky? Not enough humans to do it manually.
Now imagine the system comes back online. Does it know how to handle a crisis scenario where you have dozens of planes overhead, each about to run out of fuel? Hopefully someone thought of that edge case.