You are missing the larger issue. NNs etc cannot do their work without all the right data which Musk nor anyone using public shadow/safety driving will ever get close to collecting. Nor testing and retesting. The answer is aerospace/DoD simulation technology. (And this process kills people for no reason. Tell me what will happen when Tesla or anyone else tries to learn accident scenarios? Thousands of them thousands of times each.)
It is impossible to drive the one trillion miles or spend over $300B to stumble and restumble on all the scenarios necessary to complete the effort. Many of which are accident scenarios no one will want you to run once let alone thousands of times. Also handover cannot be made safe for most complex scenarios, by any monitoring and notification system, because they cannot provide the time to regain proper situational awareness and do the right thing the right way.
I put forth a detailed case here. Would appreciate a detailed response not just hyperbole
SAE Autonomous Vehicle Engineering Magazine — End Public Shadow/Safety Driving
Autonomous Vehicle Engineering - January 2019
Experience our interactive, profoundly engaging digital publication!
Common Misconceptions about Aerospace/DoD/FAA Simulation for Autonomous Vehicles