This is true. But Waymo is dealing with the same kind of challenges and it seems that they are already at a point, where their technology (enabled, among other

This is true. But Waymo is dealing with the same kind of challenges and it seems that they are already at a point, where their technology (enabled, among other things, by collecting tons of data) is useful. I expect Robotics (in particular, solving manipulation) to follow the same route: super expensive R&D, multiple false starts, but eventually (tons of data => breakthrough to get smarter robots).