I agree. Machine learning is not a core-competency of a typical car company and, when self-driving is ready, customers will soon stop differentiating on the saf

I agree. Machine learning is not a core-competency of a typical car company and, when self-driving is ready, customers will soon stop differentiating on the safety levels just like with airbags, ABS, or tyre treads. Either it will be 'safe enough' (and safer than a human) or it won't be for sale. It will be a commodity. Right now we have dozens of companies spending massive amounts of money on approaching the same problem in the same way.