error than if one person dies due to computer error.

As for who dies, different people would die in different circumstances constantly in all aspects of our world. It’s a weird thought that we should feel uneasy about self driving cars because of it. Like, I had a friend who died from a drunk driver crashing into him. But if he hadn’t been driving that night someone else might or might not have died. If they were both on self driving cars it’s highly unlikely there would have been an accident. But tomorrow if I’m driving along and someone does something stupid at an intersection I could die. Or next time I’m in San Francisco I could be an extremely unlucky person who is a very unlikely self driving car accident that kills me. In the end in aggregate the amount of deaths by self driving cars is a fraction of what it is for human drivers.

Also, where do you get the idea that a good human driver is better than a self driving car? Because if we’re talking about the extensively tested and eventually dispatched ones like what’s in SF, human drivers are much worse drivers. I know the statistics on this.