A big problem is that robotaxis and FSD are not the same product. The technologies are related, but they're not even the same technology. They don't have the same goals or the same constraints. For example, everyone developing robotaxis intensively 3D maps the areas in which they're intended to operate. They use lidar and other kinds of sensors in addition to optical cameras for both mapping and to operate their robotaxis. If it turns out that camera-only autonomous driving is insufficient, not only does Tesla not have the right sensors on their vehicles, they haven't got the mapping data either. Announcing a robo taxi is as much an act of creating a product by proclamation as a Mars colony.