You can disagree all you want but is the truth, AP might have save you many times but it also tried to kill me more than I would like to. One time didn’t read a line and drove me against the freeway divider.
Both scenarios can be true. During the same drive. Repeatedly.
I was refuting Jeffro01's claim that there is no way AP could ever run into a barrier so we should throw it out completely as a possibility in this crash. Just an example of AP running someone into a barrier a year ago. And considering that AP2 seemed more unstable than AP1 up until 10.4 (at least in my experience with both), I don't think we could rule out AP2 doing this either.
AP1 and AP2 have both had their share of drifting but I’ve found it worse with AP2 in that the latter will routinely cross over lane markers. It’s usually no big deal if one is in a lane that’s not a few inches or at best a foot from a string of concrete K-rails.
It seems like every AP involved accident (except the Florida one) - firetruck, China, barriers, Mountain View - involves AP in the left lane, or the fast lane. I'm going to stay away from the left lane when using AP
Well, remember back in the day, while in the rightmost lane, AP was quite happy to exit you from the freeway when you didn’t want or expect to. Most of the time there was even an exit lane, fortunately. There was one time coming toward the Twin Falls, Idaho SC when AP decided it would take that right/exit lane while in the middle of crossing an overpass. That was... not fun.
With regard to this tragic event, it makes sense that a daily commuter would have AP engaged, and we all test new versions, no matter how squirrelly the previous version was - in part because of the abject paucity of release notes most of the time, and in part because that’s what engineers and other early adopters do.
Simply put, we are a curious bunch. In more ways than one, perhaps.
What’s unclear is whether AP was engaged, manually disengaged, or programmatically disengaged *at the time of impact*. If AP locked onto the left solid line prior to impact, that would have been not good, obviously. Was the driver distracted or blinded? Did another car encroach on his lane? There are dozens and dozens of scenarios and permutations. Re-creating the impact itself is not going to be nearly as difficult as working through what really happened during the pre-impact timeframe.
I feel badly for the family and Mr. Huang’s coworkers. I hope one of the outcomes of this accident is 100% barrier reset or effective alternatives in place within 24 hours of similar accidents. Nationally. Seems the NTSB could help with that.
Jagged hardened steel pointy things pointed in my direction without effective countermeasures in place as I’m headed toward them at high speed? Not acceptable. Not a single flashing barricade? Hindsight is 20/20, sure - but this defies credulity. Multiple accidents happened exactly there previously.
I find myself thinking that the sooner we get through the next 3-5 years until decent FSD, and through the next 15 or so years after that until the majority of vehicles on the road have that technology, the better.