Ross Douthat interviews Andrew Miller, transportation policy writer and co-author of The End of Driving, about the rapidly approaching reality of autonomous vehicles in American cities.

The conversation explores the timeline for widespread adoption (projected for 2035), current safety records of companies like Waymo versus Tesla, and the complex liability issues that could determine success or failure.

Miller presents both optimistic and pessimistic scenarios for how self-driving technology could reshape urban transportation, car ownership patterns, and American culture around driving and independence.

The Safety Case and Timeline for Autonomous Vehicles

40,000 Americans die annually in road incidents, with the vast majority caused by driver error, making self-driving cars a potential life-saving technology at scale

By 2035, normal North American cities will likely have large fleets of self-driving taxis, with Waymo announcing plans for more than 15 cities after rapid expansion

The AI revolution is accelerating development because teaching cars to drive is expensive initially but very cheap to copy once solved

Bad weather poses challenges that different technology approaches handle differently - Waymo uses multiple sensors (cameras, radar, LIDAR) while Tesla relies primarily on cameras

Current Technology Performance and Limitations

Waymo operates without safety drivers and has proven safer than human drivers based on California transparency data, while Tesla still requires frequent safety interventions in Austin

Tesla's current autopilot requires constant human monitoring: "You have to maintain yourself in a state of cat-like readiness" - Miller, describing his personal Tesla experience

Waymo uses remote assistance where human operators provide guidance when cars encounter confusing situations, like traffic cones, rather than directly controlling the vehicle

Self-driving cars can experience hallucination-like behaviors, with Tesla videos showing cars suddenly veering off roads for inexplicable reasons

The Liability Challenge That Could Make or Break the Industry

Liability is "the single issue that is most in need of clarity" because manufacturers' reluctance to accept full responsibility could hold back the entire sector

One incident ended Cruise (General Motors' self-driving division) when a car dragged an injured pedestrian after hitting them, showing how single accidents can destroy companies

Regulators hold self-driving cars to much higher standards than human drivers, requiring "six nines" (99.9999%) safety rather than just better-than-average performance

Waymo has accepted liability for their current fleet, while Tesla has been reluctant to assume responsibility for their driver assist systems

Geographic Reach and Political Obstacles

Rural America won't see self-driving cars, similar to how Uber isn't viable in rural areas, but suburbs could work with public stipends to robo-taxi companies

Political obstacles scramble traditional party lines, with Democratic concerns about labor displacement and Republican concerns about surveillance and data privacy

Texas and Tennessee are more open to self-driving than blue states, creating potential for future culture war divisions over driving regulations and insurance rules

Public transit agencies represent a major obstacle due to heavy unionization, viewing robo-taxis as threats to livelihoods rather than complementary technology

Two Futures: Utopian Efficiency vs. Dystopian Congestion

The good scenario features 40-50% cost reduction, households giving up cars (two-car families become one-car, one-car families become zero-car), and reclaimed parking space for better uses

The bad scenario creates worse congestion without reducing car ownership, while public transit enters a death spiral as people defect to cheaper robo-taxis

The fork depends on regulators allowing widespread robo-taxi deployment and public transit agencies embracing rather than fighting the technology

All automated vehicles in development are electric, potentially accelerating the transition away from internal combustion engines

Cultural Loss and the End of Driving as Embodied Knowledge

Driving represents embodied knowledge and a rite of passage that Miller acknowledges will be lost: "We don't have many rites of passage for young people anymore"

Miller argues the benefits outweigh cultural losses: "If we have a tool that can save lives while also giving people their time back, I think we would be a fool not to pick it up"

Insurance costs could eventually price out human driving as self-driving becomes exponentially safer, creating gradual but inexorable pressure toward automation

The trend fits broader American screenification and risk-aversion, with teenagers already postponing driver's licenses in the iPhone era