A 2025 Tesla Model 3 in Full-Self Driving mode drives off of a rural road, clips a tree, loses a tire, flips over, and comes to rest on its roof. Luckily, the driver is alive and well, able to post about it on social media.
I just don’t see how this technology could possibly be ready to power an autonomous taxi service by the end of next week.
This represents the danger of expecting driver override to avoid accidents. If the driver has to be prepared enough to take control in an accident like this AT ALL TIMES, then the driver is required to be more engaged then they would be if they were just driving manually, because they have to be constantly anticipating not just what other hazards (drivers, pedestrians,…) might be doing, they have to be anticipating in what ways their own vehicle may be trying to kill them.
Absolutely.
I’ve got a car with level 2 automation, and after using it for a few months, I can say that it works really well, but you still need to be engaged to drive the car.
What it is good at… Maintaining lanes, even in tricky situation with poor paint/markings. Maintaining speed and distance from the car in front of you.
What it is not good at… Tricky traffic, congestion, or sudden stops. Lang changes. Accounting for cars coming up behind you. Avoiding road hazards.
I use it mostly like an autopilot. The car takes some of the monotonous workload out of driving, which allows me to move my focus from driving the car to observing traffic, other drivers, and road conditions.