When a driverless car can't handle the situation

[Ulmo]

Kinda like a military drone pilot.

That's what I was going to post: not only has this been done before often and current autonomous aircraft developers are planning this in their suites, but I believe plenty of smart people have thought of this. The problem with cars is that you can't always connect to them in every situation: satellite only works when the car isn't under anything or behind something, and local radio can't get everywhere. Another thing is that on the ground, things are already touching (people cling to ground with something called gravity), so it's harder to react in time, whereas in the air, there's often more reaction time since most things are far away at first.

The ferry one does sound like a difficult task, but I'd argue that's not a driving task that has to be solved with level 5 since you're transporting the vehicle instead of the vehicle actually driving. I realize there are some places only accessible by ferry so I'm sure there will be a solution in the future but I wouldn't expect it in an initial rollout.

That's pretty funny to me. I've never seen cars on a ferry except in weird foreign videos. Bjorn just made one. It was the most odd thing! So, I suppose it does exist. But it doesn't really seem like the natural world to me. I wouldn't be surprised if other developers view ferries in the same way I do, and that support comes at a later time. But, having said that, it's just another AI problem that will be easily solved once the problem is presented to the learner, if they can get access to a situation in which a ferry is available.

 

[electracity]

Confirmation that my idea in post #1 is being considered by Nissan:
Nissan plans to make robot cars; human 'mobility managers' will intervene when needed

Google will first simply move the guy currently sitting in the car to a remote location. Early Beta will be one engineer monitoring one car. Later google taxi will call for help when confused.

Tesla is probably disadvantaged at the beta stage of autonomous driving. Google will have expensive equipment in the car, including inside cameras and redundant communications.

Tesla Network using privately owned S/X/3 should probably be considered the best case scenario. Tesla may need to branch AP2 development and run corporate owned M3 with more advanced, extensive, and expensive hardware to reach level 5 and allow remote supervision of beta Tesla Network services.