alt Hacker News> Even if it prevents like ten deaths, the LiDAR will have paid for itself.
The statistical value of a human life (for US government purposes) is between 7 to 14 million USD depending on department etc. (Unless there's anything to do with nuclear, then it's pretty much infinity.)
I'm not quite sure how much LiDAR would cost extra?
Of course, in addition to lives saved, you have other benefits. Eg you can conceivably drive more aggressively and faster, when you have more scanners, because you can exploit more situations where a lesser system has to act conservative when it can't be quite sure.
It would also just be good PR at this point. They could easily spin it as "We could make it work like the others with only cameras, now we add back more sensors to improve even further!". Camera-only has been criticized prominently for so long too, it would undermine a lot of the negative arguments people make too. It would honestly just be an easy win for them, even if there is no real improvement at first, they can even take their time to update the software over time.