I think CR isn't out-of-line with their criticism, but I just disagree with it. There's no substitute for real world traffic data, and giving owners the option to be guinea pigs seems to be the perfect way to do it. Since autopilot has to be actively engaged, and the driver (via personal driving profiles) knows what he's opted in/out of, there is no additional risk from doing it this way. After all, a speeding/drunk driver is subjecting other drivers to much more involuntary risk than a non-speeding autopilot that occasionally misses stop signs (I'm not trying to be cavalier here).DougWantsALeaf wrote: ↑Fri Sep 04, 2020 11:50 amOk, I know I am asking for some potentially firey responses here, but what do the Tesla drivers here think of consuner reports review of full selfdriving? Does this reflect your experience? I only know of propilot and its limited use cases.
When I engage autopilot on city streets, I become hyper-vigilant, because I'm actively choosing to add data to the system, not to become another statistic. I have never engaged autopilot on city streets unless it's to test the system. All my city driving is fairly short, and the current iteration of autopilot will disengage when you leave the highway, so there hasn't been a good opportunity to take advantage of autopilot for city driving. But giving the drivers the option to train the system, that's actually a smart move.
So is it worth $8000? Only if you don't want to pay $10,000 later when the robotaxi system goes online. The system is constantly collecting real world traffic data to better train autopilot, and constantly improves because of the amount of data that Tesla has access to. It's just a matter of "when" the system becomes capable enough to be truly FSD.
Can you live with basic autopilot only? Absolutely, basic autopilot is pretty capable as is, so choosing to forgo FSD can also be a good choice.