Yes, if there's any chance in hell of self-driving becoming feasible at scale it will involve pre-defined routes, possibly with other sorts of monitoring systems, and new infrastructure/mandatory equipment for safe pedestrian crossing zones after the first handful of school kids inevitably get plowed down thanks to the obviously-not-quite-there-yet image recognition systems.

Likely we'd see some rollback to a more achievable goal of a city-funded fleet of robo-taxis running pre-defined routes with standardized equipment and maybe some years into it, when we realize traffic jams still suck, start thinking "hmmm.... maybe we should've just improved our bus/rail systems..."

Cities are just as easily duped by guys like Elon Musk as any of these poor fuckers who died actually entrusting their lives to their shitty "autopilot" system. Especially when cities/officials stand to profit from kickbacks of various sorts. Don't assume something like this won't come at the cost of not investing in the obvious competing tech: public transport.

They're already in conflict everywhere. Infrastructure for cars robs public transit infrastructure blind in lots of government budgets. The only public transit category potentially benefiting from car infrastructure is buses, which are arguably the worst form of public transit to begin with, and still also require additional dedicated infrastructure to get any better (e.g. dedicated bus lanes).

"Self-driving" cars obviously require car infrastructure which already steals from public transit budgets both federally and locally, but if we add government emphasis on this technology and start to develop specific infrastructure for "self-driving" cars (walled off routes, communications appliances, etc.) then they'll start taking even more of the budget.

And all of this for something that's arguably much more braindead and useless and consuming of R&D dollars than the obviously more efficient, already technically possible forms of transit that could be built or expanded upon today.

Yes