My point, was that Tesla missed something that could become an issue. GM missed something too (although I'd argue that while they should just detect if the car is actually moving, it's still a user error), so we should stop acting like GM are a bunch of idiots all because they don't make an EV on par with Tesla (which is really what all the hate comes down to).
Sorry to nitpick, but it is not a user error, but rather a 'use error' - meaning a designer needs to understand normal user behavior and build that into the design.
For instance, if you had new Microsoft software & performed a normal 'copy and paste' series of keystrokes, you'd be very upset if it erased your entire hard drive without warning. And I doubt you'd find it okay if you called Microsoft and they answered, "Well, you should have read the manual, it was in there." (Actually, you'd have a right to be irate at that point, because you EXPECT certain actions to happen.) So with GM, it is not an unreasonable expectation that a car does not start especially in a closed garage. That's a use case that just should not happen. Ever.
'Use error' doesn't assign blame, 'user error' does. (Welcome to the world of medical device design, where you need to factor in the user(s) expectations of behavior.)
/endofnitpick
- - - Updated - - -
In the case of Tesla and road debris, that may very likely have been a known outcome. Typically a risk analysis would be performed. If they understood the failure mode/severity (car notifies driver, gives plenty of time to pull over and get out), and believed the probability was low, they may have decided to live with the outcome.
When the probability changed over a short period of time, they made a modification to change the probability of the failure occurring.
In the case of GM, if they had identified 'car may start unattended in a closed environment', the severity would have been high enough (potential death) that no matter how low the probability was, it would require a design mitigation.
Make sense?