Idiot of the year contender

[eddie]

12 minutes ago, Moist One said:

the problem you have is, you would need <insert HUGE number> sensors and <insert huge number> processors to match what the human eye and brain can see and interpret. You would probably also need a few miles of fibre optic cable, and a variety of different voltage cable, and probably more code than @eddie has seen in his entire career just to complete and read the circuit of signals required just to move the car.

If you look at Tesla, and read all their bad news, they cannot launch cars on time simply because the software isn't ready.

Even something as seemingly straightforward as lane finding and following is vastly complex in practice. The code had been under development for a decade or more - of course it's not just receiving input from optical sensors and then reacting, because as soon as the vehicle has reacted (and its position moved), there is more input, more re-positioning, more input - not just the lanes but vehicles in the lanes (are they acting in a predictable manner or not?), vehicles that are following, to the side, not moving (breakdowns - is the entire road blocked or just this lane, can I overtake safely or is there a vehicle already overtaking me), are there road signs that are giving instructions, is there a vehicle doing something it shouldn't, e.g. coming towards me from the other direction etc etc.

We're talking thousands of objects (programs), tens if not hundreds of millions of lines of code. There is self-driving software (open source) on GitHub written in Python - it's been years under development with an awful lot of people contributing their time.