I disagree. I'm not at all trying to dismiss the significance of the problem; apologies that I didn't communicate that effectively. The systems absolutely SHOULD get better and be better than us. However, I think that comparing the difference to human perception helps enable us to discover differences in how these systems view and process the world around them and gives us hints at how to improve them. After all, humans also have these problems to some extent, and knowing how much better or worse AI is at it can either help improve safety for human drivers too or maybe highlight systemic issues that need to be addressed outside of the vehicle (e.g., daylighting, improved crosswalk infrastructure, et cetera).

Maybe not worth it for you, and that's fine!

Cool idea, but I think long press mostly filled that niche already.