> In his influential 2011 book The Lean Startup, Eric Ries advises companies entering a new market to develop a "minimum viable product"—the simplest product go

> In his influential 2011 book The Lean Startup, Eric Ries advises companies entering a new market to develop a "minimum viable product"—the simplest product good enough to attract paying customers. This strategy allows a company to begin selling real products to real customers—and get real market feedback—as quickly as possible. I think the problem is that they're only interested in selling fully autonomous self-driving vehicles. There's very little market advantage in selling a "mostly" self-driving car, and Waymo doesn't want to be in generic car sales for years until they're ready. So they have to hold out for the 0 to 100 switch flip. And they don't get to make mistakes and iterate on core functionality, so they have to wait. People's lives and schedules will be at the mercy of these vehicles, in a wide variety of operating contexts. They have to get it 100% right. If they kill people the brand itself is potentially dead. So the picture is they need to go 0 to 100 with near 100% accuracy. Anything less is at best brand dilution or at worst corporate suicide. Tesla's approach matches the MVP idea more closely since they can add small features incrementally. Presumably their eventual goal is to be able to say "with this update your car is now fully autonomous". But they're in the business of selling cars to begin with, so every release short of full autonomy is A-OK. Autonomy is a feature, not the product.