Self-driving cars still 25 years off in the USA? No.
Bill Gurley argues on CNBC that we are 25 years away from autonomous vehicle market penetration in the USA because we’re too litigation-hungry. And concludes that AVs will instead take hold in a country like China which has relatively uncrowded roads and an authoritarian government that can make central planning decisions.
I don’t agree. Precisely because of rampant litigation in the USA, insurers are going to do the cold, hard math (like they always do), and realize that AVs will save a passel of lives and hence be good for their book. They will therefore indemnify manufacturers or otherwise shield them from opportunistic lawsuits launched in the inevitable few cases where the cars are apparently at fault. Money will smooth the path to AV adoption.
He also says:
The part we haven’t figured out yet, the last 3 percent, which is snow, rain, all the really, really hard stuff — it really is hard. They have done all the easy stuff.
While I would agree that there are still some really, really hard things to work out in AVs, rain and snow aren’t among them. Sensors like radar can penetrate that stuff far more effectively than human eyesight. Even pattern recognition in the optical spectrum could outperform humans.
The hard part part is getting the cars to know when they can break the rules. A recent viral posting about how to trap AVs hints at that. When a trash truck is taking up your lane making stops and you need to cross a double yellow to get around it, will an AV be smart enough to do that? Sure, it can just sit there and let the human take manual control, but that doesn’t get us to the Uber-utopia of cars making their way unmanned around the city to their next pickup.
