You don't need as much ventilation since there are no products of combustion to exhaust, and the risk of crashes from human error are eliminated. Presumably the car software and the tunnels themselves can also be designed in a way where the risk of crash is near-zero.
Public transport is absolutely gross and legitimately scary in Los Angeles. However, if I could stay in my car I would totally use this system, also having your car at your destination is incredibly useful and not something you can do with a subway.
IMO they both serve different needs.
And if you're talking about FSD, most of those instances are from exceptions eg. unexpected pedestrian, parked truck, construction works.
Its fair to imagine that these tunnels will be custom-built for self-driving cars and will be totally controlled environments.
Toyota runaway acceleration caused by a stack overflow: https://en.wikipedia.org/wiki/2009%E2%80%932011_Toyota_vehic...
Driver error or pedal misapplication was found responsible for most of the incidents. The report ended stating, "Our conclusion is Toyota's problems were mechanical, not electrical." This included sticking accelerator pedals, and pedals caught under floor mats."
"The cost of loop stations is so cheap that you can place them next to any number of desirable locations."
This also feels wildly handwavey. You need a LOT of them if you want this to be a viable point to point network vs just a higher-risk way to build a subway. With individual vehicles going to individual stations at user request, too, we seem to have brought in "full autonomous driving" as a prereq.