I don't like Uber but as far as autonomous driving is involved, Uber and Waymo use Lidar AND camera arrays versus Tesla only using camera arrays. I don't see at all how that is a deal breaker.
Just because humans can operate a car with vision-only capabilities doesn't mean that we shouldn't use LIDAR. Every redundancy will be highly valued when human lives are at stake, and we've seen many times with Tesla how cameras alone can be fooled.
If you are pursuing the Tesla self driving car business model which is selling a mid range luxury self driving car directly to the end consumer (and potentially operated for 4-6 hours a day as an autonomous taxi) then the cost of sensors is paramount.
If you are pursuing other business models such as Uber and Lyft and Waymo, where the taxi will be purchased as a fleet, fleet managed by the operator and operated 24x7 for 5 years straight, then the cost of camera sensors + LIDAR suite will be full amortized over the 5 year life of the car and therefore be somewhat of a rounding error in the total investment return.
To admit otherwise would be disastrous for stock price and future sales. And to include lidarr at their current sale price would not be possible.
It will beat comparable solutions that contain that expensive component.