I would agree that it will set back the public perception of it as "safe", but I'm not sure that if it is found that there was driver error (beyond improper use), that it will slow "the industry" down. The industry (Tesla in this case) is poring over all the data as we read this thread, and likely has the "problem" narrowed down. So then, the public perception has probably suffered a greater setback than anything. I LOVE TO DRIVE, and I think it would be hard NOT to drive a car like a Tesla w/ no autopilot (I mean aside from the range anxiety that would surely fade). There are applications where an autonomous car (or a car anywhere along the continuum of "autonomous") would work wonderfully, but as many have pointed out humans get inattentive, so accidents are likely still going to happen.
I'd like to see data on average accidents per mile on autonomous vehicles vs average accidents per mile on "regular" vehicles. Problem is that if the comparison showed that autonomous cars were safer or "as safe" as regular ones, then humans would likely get even MORE inattentive/cavalier than they already are.
This Sky's not perfect, but it's perfect...for me