Ashok is starting to talk more about v15 here. The conversations I had at the Cybercab Launch day were in much the same context. v15 has some new "functions" that will enable a safer and more reliable path to autonomy. In this post he mentions earlier prediction time .. to me this means more context length, maybe even better vision to voxel to occupancy network. He also mentions faster reaction time which implies an even more optimized photons in control out pipeline. All very impressive work by the @Tesla_AI team, and I can't wait to see it first hand.