This is purely my point of view about Ford’s usage of lidars which I believe will slow them :
Tesla’s unconventional approach to self driving through solving vision (and thus not needing lidars) is very pragmatic actually.
Lidars will provide you 3D meshes that you will be obliged to analyse and comprehend.
If you train your neural network to actually very accurately perceive 2D images, you are already ahead of lidars which cannot yet fully make sense of what they see. Training neural networks on lidars is much more complicated.
I think Tesla’s approach by solving vision is based on the idea of avoiding Lidar’s local maximum.
If you can perceive well enough in 2D (safe enough for understanding your environment accurately), you won’t need 3D perception anyways.
Plus if you can’t handle 2D perception well enough in the first place you wouldn’t be able to handle 3D perception neither since it’s a lot harder to master.
Ford didn’t quite think about it this way.
Let’s not forget that this isn’t a race of economics but safety. People die on the roads because of other people’s stupid mistakes. If one could simply solve this, you would save more lives than ending civil wars..
Tesla’s approach of simply using cameras and developing cost efficient custom chips is essentially solving accessiblity. Waymo may be cool but the equipment there is much more than what any average consumer could afford. (Yes Tesla isn’t cheap, yet, neither)
Disclaimer : I am no expert, just giving my perspective from a student engineer’s perspective. Feel free to comment and argue