It’s bizarre that people tried to disprove the safety test by rerunning it and showing that sometimes the car will brake. Clearly this is a case where a single failure can nullify any number of successful tests. The problem is that the car will hit a mannequin sometimes, and it doesn’t matter that sometimes it won’t.
The logic behind this entire experiment is flawed. My 2012 jeep won’t just hit a mannequin. It will hit literally everything unless I intervene and stop it. But since a Tesla will stop itself for some things, it must stop itself for all things? Progress happens incrementally or not at all. We went from cars not stopping themselves for anything to cars sometimes stopping themselves and people are upset?
You still have to drive. Anyone with the beta knows this. The thing literally beeps at you if you take your hands off the wheel or look away for too long. This “machines will kill us all” hysteria happens with all new technologies. Humans used to crank elevators by hand. Sometimes they messed up and cut people in half. So we had machines do it and naturally hysteria ensued. Seems to have worked out though.
I'm not acquainted with how Tesla cars do object detection. Is thermal imaging involved? Could a mannequin be a problem because it is not hot like a human?
