I'm saying that it was probably active until the driver disengaged it to make that lane change. Everything about that video says "confused driver" and not autopilot misbehavior. The car will let the blinkers run for 3-4 seconds before a lane change, and it's very cautious about when to do it (much more so than a human driver would be).
Do I have proof? Of course not, it's not my car. But that's what it looks like. And I'm pointing out that this is exactly what happened in the accident in the linked article, where Tesla drivers were all pointing out how this just didn't match AP behavior[1]. And they were ignored. And they were right. Just wait.
[1] In that case: the version of autopilot at the time would only engage on roads with lane markings, and this road didn't have them. Also the speed of the collision could only be reached on that tiny road with manually-commanded acceleration, AP is quite timid.
Comments
Merely stating it was active _within_ 30 seconds, that doesn’t mean it was deactivated at the 30 second mark. Or am I missing something?
I'm saying that it was probably active until the driver disengaged it to make that lane change. Everything about that video says "confused driver" and not autopilot misbehavior. The car will let the blinkers run for 3-4 seconds before a lane change, and it's very cautious about when to do it (much more so than a human driver would be).
Do I have proof? Of course not, it's not my car. But that's what it looks like. And I'm pointing out that this is exactly what happened in the accident in the linked article, where Tesla drivers were all pointing out how this just didn't match AP behavior[1]. And they were ignored. And they were right. Just wait.
[1] In that case: the version of autopilot at the time would only engage on roads with lane markings, and this road didn't have them. Also the speed of the collision could only be reached on that tiny road with manually-commanded acceleration, AP is quite timid.