I disagree with the click-bait comment. I thought this was one of the best written criticisms of the current path to self driving cars I’ve read. While I disagree with the conclusion, they presented all the facts clearly. They didn’t just focus on Tesla and Ubers bad track records but also included where Google and GM are on the path. I think their criticisms of deep learning and how the industry is working around them are spot on.
While every point they raised were good ones, I just disagree with how difficult they are to work around. The rules based glue to deep learning models is a sound one and I don’t even see it as a negative. The industries reliance on detailed maps makes self driving a much more rules based problem. Hopefully the level of detail on maps can be relaxed over time but to start off that way isn’t a deal breaker to me.