Agreed; I imagine they use neural networks to detect and classify objects which are then saved into a scene-graph for use in pathing.

I expect what happened was that they trained their NNs for improved detection in one area but unknowingly reduced it in another. Perhaps now it can detect tricycles 99% but road barriers went down to only 30%. Having worked with NNs it's very common to see gains in one domain which come at a cost of reduced performance in another.