There is a reason we put age, training and health limitations on human driving though.
Yes, that's why 85-year-old cardiac patients with cataracts are never allowed to drive without a redundant driver sitting right next to them.
It's also why young drivers are not allowed on highways with speed limits above 45 mph until they turn 18, have passed a 40-hour driving instruction school and are not on any kind of medications (like for allergies or depression). The USA is especially strict about this due to the lack of requirements for a redundant driver.
And they have a superior brain, the human brain.
The capabilities of a FSD car, ready for widespread deployment, has not yet been demonstrated so it remains to be seen what its capabilities will be. But you sure seem to have a high opinion about the functioning of human brains. I have many years as a motorist under my belt and my observations have not been so kind. In many cases, the most rudimentary machine could do better. I have seen human brains do the most assinine things. I see the hurdle as a very low bar.
The question is not will redundancy be needed. It obviously is and even Tesla has some. The question is what kind of redundancy and technology will be sufficient for autonomous driving — and further what is best for that task.
Agreed. Redundancy is a good thing, to a point. It's possible to take it too far. The bottom line will be demonstrating an overall level of failure (odds of injury/death crashes) that is significantly lower than human drivers achieve. I tend to believe that redundancy won't be regulated directly, the presence or lack of redundancy will be "built-in" to the safety statistics that are used to inform us whether FSD is more or less safe than human drivers.
Tesla is different from the rest of the industry in that they have made a very clear bet on what this answer is — and is not. They are shipping and selling a suite they say is Level 5 capable hardware without geofences, and they have voiced their opposition to one technology in particular (Lidar).
Tesla is not opposed to anyone else using whatever technology gets them to FSD most quickly, but they have made an informed decision that it's more trouble than it's worth and Musk predicted that others development efforts based on LIDAR will abandon the technology, not due to any inherent opposition to the use of LIDAR but for practical/technical reasons.
As far as I can see everyone else still says we don’t know what the final technology requirement will be. Even Waymo responded in that kind. I would not be surprised if Tesla too ends up changing and expanding their suite down the road.
Again, I don't think regulators will require any particular technology, it will be the capabilities of the system that will be regulated and this will be informed by demonstrating the system is safe and effective over millions of miles. If it is not safer than humans with a high level of confidence, it won't be allowed. As FSD systems mature, and there are multiple solutions in use, the systems will need to show they are as safe as other competing systems. In other words, the safest FSD systems will, over time, raise the bar for other FSD systems. Statistics will be used to determine which systems are safe enough to continue to be approved for sale.
