I'm still surprised that they, and most of their competitors, chose to start off in such difficult to navigate regions. I would have started off in a retirement community or the like.
Not operating in difficult areas would slow them down because progress in this domain is bound by the breadth of edge cases encountered in the wild.
1. Silicon Valley, both for the workers and the general public's willingness to embrace fancy new tech 2. SF is DIFFICULT to drive in. If you follow all driving laws & drive super defensively, you'll never get anywhere on time. If an AI can learn to effectively navigate around traffic, construction, stopped rideshare and foodshare cars, large crowds, etc etc, it can drive anywhere.
SF also gets a lot of rain and dense fog, good for testing low-visibility conditions
I think you are really understating it. I believe Chandler, AZ has been Waymo's proving grounds for years, and according to this article (https://www.wsj.com/us-news/waymo-phoenix-arizona-self-drivi...) the Phoenix area is the leader of autonomous vehicle service.
It’s just absolutely normalized here to the point of barely being a novelty anymore. Waymos are absolutely everywhere.
Also, the risk aversion of people in retirement zones does not make it a great target customer base.