automatic driving cars are not going to become the norm for a long time because they're unpredictable. many fail and crash.
They actually don't, believe it or not. Google's self-drivers now work well enough to cause significantly
less accidents than human drivers. If you read their
monthly reports, you'll find that a common case is where a self-driving car will predict something from another driver, but the human driver does something differently.
To put things in perspective, self-driving cars are logically more likely to be safe than human drivers, simply because computers always do what we tell them to do, provided there's power and all that. Humans can be tired and inattentive, and are prone to making tons more mistakes and faults in judgement when driving. If you type
print("Hello World") it'll always print Hello World. If a car detects another car in front of it and is instructed to stop, the car will always stop, most likely at a rate depending on its current speed, distance and all that. With that in mind, if
every car were to become autonomous, each car could then predict and coordinate the actions of every other car with near impeccable accuracy, and safety wouldn't be an issue anymore. Any flukes would always be a case of an outside unexpected influence.
Of course, there's the human error in programming the cars themselves, but that's what testing and prototyping is for. Once all the edge cases are all buffed out, you have a safe, viable product for consumer use. And I think it's beautiful.
