I don't see why one of those second-order effects wouldn't be the death of car ownership, with everyone using a rideshare service instead. Hell, that's the business model for Waymo and almost everyone other than Tesla in the autonomous-vehicle industry. It just doesn't make sense to own your own vehicle, use it for ~2 hours/day, and have to worry about parking/storing/fueling/maintaining it when you could have a service do all of that for you. Plus self-driving cars fix several issues with human rideshares, eg. you can drive it out to the boonies without worrying about how it's going to get back; you don't need to worry about getting assaulted by the robot driver; when they wait for you you only need to pay the opportunity cost of another ride rather than the opportunity cost of the driver's time. It's feasible to take a Waymo out to a state park, though you wouldn't usually do that with an Uber.
The second-order effects of that could be pretty wild. If people stopped owning their own cars, we wouldn't need houses with garages and driveways. It'd favor dense development with loading zones rather than parking spaces. It'd also be a big boon for EV adoption since the cars are all owned by one corporate owner and all go home to a centralized depot to charge at night rather than needing to retrofit EV chargers onto everyone's living situation. (Indeed, Waymo runs an all-electric fleet.). There'd be a premium on very reliable powertrains, since the cars might easily put 60-70K miles/year on them instead of the 10-15K that is typical of passenger vehicles. I dunno why Waymo went with Jaguar instead of Toyota, but perhaps "EV" is the explanation. Cars would wear out in 3-5 years instead of lasting for 15-20, and so you'd always have the latest hardware and technology on the car.
All the money we spend on traffic enforcement would become pointless, with audits of the software becoming a more effective use of dollars instead. But that blows a hole in many small local PD's budgets, many of which use speeding and parking tickets to raise revenue. Municipalities would likely find themselves powerless at regulating Big Self-Driving Rideshares.
The third-order effects are interesting as well. Once all cars on the road are self-driving, why not have them draft each other and physically link up to improve power efficiency and safety? You might even call such an arrangement a "train", blurring the line between road and rail transportation. But then, if you've got docking and linkage mechanisms, why not put the boundary between the electronics & powertrain and the passenger compartment, like the Rivian "skateboard" platform? You could return to private ownership of the passenger compartment - where, after all, some people like to store all their junk - and then have the rideshare own only the means of locomotion. Then you could extend this to other forms of locomotion like elevators, airplanes and ferries, so that your passenger compartment could just drop down an elevator shoot, onto a waiting self-driving car, which links up with others to become a train, takes you to the airport where you're loaded onto a plane without ever having to board, and then your pod deplanes and a self-driving car takes you straight to your hotel, where you now have transportation to wherever you want to go.
The future looks an awful lot like intermodal containers for people.