To me it's such a strange argument. Why wouldn't you want CV + additional sensors? CV has made all kinds of progress in current years, but it's certainly not fool-proof, doesn't do well in low-light or extremely high-light scenarios, can be slow, difficult to debounce between frames, subject to motion blur, difficult to get depth from, subject to optical illusions, etc. There are solutions that aren't full-blown LiDAR that can do a great job of depth estimation, collision detection, etc. Wouldn't you want more sensors in the mix? To me it's a really odd 'party line' to adhere to.