Autonomous vehicles are becoming more and more commonplace everyday. But whether they’ll induce endless sprawl + super commutes or help us all finally go car free is still unclear. To find the answer to those questions, I spoke with my good friend and autonomous transportation expert Andrew Miller — enjoy.
Andrew Miller and The End of Driving
[00:00 - 04:03] Andrew Miller, is the author of the Substack Changing Lanes as well as the new book, The End of Driving: Automated Cars, Sharing vs Owning, and the Future of Mobility.
The End of Driving presents the following theses:
We have a two possible futures with AVs: one with privately-owned vehicles (bad) and one with shared AV transportation systems (good)
We’re heading toward the bad future by default.
The Two Paths Forward
[04:03 - 06:20] Private ownership is the likely default outcome due to path dependency, established car company infrastructure, and regulatory barriers. Robotaxis face fierce municipal pushback (example: potential Boston ban), while privately-owned autonomous vehicles only require federal safety approval.
Who’s Building What?
[06:20 - 10:19] Waymo dominates shared mobility, while traditional automakers (Mercedes, GM/Cruise, Tesla) are all incrementally advancing toward private ownership through driver assistance features. ODD (Operational Design Domain) refers to the specific conditions where autonomous systems can operate safely.
Risk Profiles and Use Cases
[10:19 - 13:13] Urban robotaxis operate at lower speeds with simpler failure modes (pull to curb), while highway autonomous systems face higher stakes despite more predictable conditions. Automated trucking represents a promising but slowly-developing use case.
The Elon Musk Question
[13:35 - 19:25] Elon Musk deserves respect for achievements (SpaceX, early Tesla, PayPal) but employs deliberate hype as a strategy.
Tesla’s camera-only approach (versus competitors using LIDAR/radar/GPS) may work eventually, but proving safety to regulators will take time—potentially five+ years.
Meanwhile, Tesla faces competitive pressure from Chinese manufacturers like BYD globally and political backlash domestically. The cyber cab incident (driving into oncoming traffic on launch day in Austin) illustrates the challenges ahead.
Policy Solutions for Better Outcomes
[19:25 - 28:06] Key interventions to avoid the bad default future:
Regulatory harmonization: State-level robotaxi approval should be as straightforward as federal approval for privately-owned AVs. California’s CPUC model provides a positive example.
Super blocks for AVs: Zones where only autonomous vehicles can operate, moving at pedestrian speeds. This gives urbanists car-reduced streets while maintaining access benefits—autonomous vehicles will naturally route around these zones unless necessary.
Transit agency transformation: The most critical challenge. Rail systems with dedicated rights-of-way will survive, but bus-based systems face existential threat from robotaxis. Solutions include automating existing rail and buses, and agencies operating their own robotaxi fleets for paratransit, feeder services, and subsidized first/last-mile connections to transit hubs.
Traffic Laws and Autonomous Vehicles
[24:00 - 27:24] Traffic laws function as guidelines with “reasonable person” exceptions. Autonomous vehicles can be programmed to specific thresholds of reasonableness, allowing enforcement of stricter compliance (like slower speeds in super blocks) that can’t be reliably expected from human drivers.
Current Progress and Examples
[28:06 - 35:01] San Francisco emerges as a success story—despite initial resistance, city officials have worked to help Waymo succeed once mandated by the state. Phoenix and Miami also deserve credit. Disempowering the most vocal blockers (often city councils and mayors) while working with professional civil servants yields better outcomes.
Liability and Safety
[35:01 - 42:44] Manufacturers must accept 100% liability for autonomous vehicle accidents as both good policy and practical necessity for public adoption. The contrast between fully autonomous vehicles (Waymo/Zoox) and partially autonomous systems (Tesla autopilot) is stark—the latter creates dangerous situations where humans must stay alert during automated driving, similar to airline autopilot challenges.
What to Expect in 2026
[42:44 - 45:27] Waymo’s rapid expansion from 5 cities at year’s start to over 15 now signals accelerating deployment. Zoox is launching in Washington D.C., which could influence lawmakers. Tesla faces pressure to show progress despite readiness questions. Most people who haven’t tried a robotaxi yet will have the opportunity by end of 2026.
How to keep up with Andrew and his work
[45:27 - 46:42] Andrew directs listeners to his book “The End of Driving” and his Substack newsletter Changing Lanes, where he publishes weekly on autonomous vehicles, transit, freight, and more.
Thank you to
, , , and several other folks who tuned in live.
