I don't even like the idea of drive-by-wire controls. Give me a physical throttle cable and steering linkage any day. When it comes to self-navigating vehicles, I just don't think something as complex as safely operating a moving vehicle can be reduced to a layered series of glorified "if/then" functions.

I'm certainly no computer science expert, but I was recently pondering why sometimes I have random software issues with my computer when at other times, everything works flawlessly despite everything seemingly being the same. The only thing I could figure is exactly what this article describes - "code is piled on top of code", and perhaps the outcome of a command depends on the order in which the operations are sequenced. Perhaps there is some error in the code logic, that when a particular algorithm is applied, results in a error that otherwise doesn't occur when the same command is executed at another time.

I'm talking out of my rear end here, because I have zero understanding about how coding actually works, so any of you coders feel free to blast my feeble attempt at grasping the concept. The bottom line is that I don't trust computers to do things that require a lot of split-second decision making and could have deadly circumstances. It just seems that simulating complex processes that are affected by nearly infinite variables with algorithms is an exercise fraught with potentially disastrous consequences, because catching the errors before they actually manifest seems to never happen. It's impossible to simulate in a lab setting every single driving scenario that might occur, so to some extent, it seems that self-driving vehicles are being beta tested in "the real world".