my friend bought a ride from a waymo for shits and giggles. i was actually intrigued about how the car made decisions on its own and sorta nerd sniped me to learn more about machine learning (not llms).
there’s probably some evil thing the self driving car companies are doing that im not aware of, but I was just interested in the ML aspect.
I think that advocates are right in that, once the kinks are worked out, they will be far safer than human drivers.
At the same time, I think the implications lead to a lot of dystopian implications for cities. Like, a properly tuned self driving car will never hit a pedestrian. So pedestrians in cities will learn they can safely step out into the street even in oncoming traffic. But this will cause the car to swerve and slam to a halt, which us a bad UX for the passenger. Enter: automatically reported jaywalking offenses powered by car cameras and facial recognition technology.
The kinks will never get worked out - just enough so that the payouts they have to inevitably make don’t eat into their profits too much. I don’t say that as “this technology sucks” so much as “this is a problem without a solution”. Roads, traffic, pedestrians, etc are all too random to design self driving vehicles that don’t make mistakes. We could have gold plated public transit and every sort of pedestrian friendly bell and whistle for the amount of funding it would take to get 80% of the way there. And it doesn’t actually solve any of our infrastructure problems, it just adds “convenience”.