You came back pretty forceful here using emotional scenarios to try to force your point. I am not sure why. Does Musk's association with Trump generate that much negative emotion?

Vision is proven to be the superior system, in every way, its more vibration proof, less expensive, and it does far more, if the latest technology in it is available, as explained.

For redundancy / backup, more of the same can be added, and will have been added, there will for sure be multiple systems. This was a sufficient analysis, and actually demonstration, for Aerospace. Why would it not be for cars? You get that all the cutting edge techniques of software, testing, safety risk assessments, safety standards etc come from aerospace, to cars, not the other way round, right?

FSD won't be perfect, no matter what form it comes in, there will be accidents, including fatal. But the frequency of accidents is what the various systems will expose in the end. If it's better than the manual case, which wouldn't be difficult, the accident rate will probably be accepted. This was the criteria sought by CAA, FAA, etc.

If the vision systems have motion capture capability then they can not only predict collisions with pedestrians, but also collisions with things, or even other humans or animals the pedestrian might throw in the path of the car.

There is no other system that can do that.

And I hate to say this, but it would be a simple and spectacular demonstration, in perfect character of Musk, to show the vision system avoiding something like that, varied every time, avoiding it every time, whilst the less equipped cars incapable of detecting it, would drive into it, every time.

Then the arguments would be over, right?