The future of driving is autonomous with Dmitri Dolgov of Waymo

Episode Summary

Executive Summary: Guy Raz interviews Waymo co-CEO Dmitry Dolgov about the company’s path from DARPA challenges and Google’s secret self-driving project to today’s autonomous ride-hailing service in San Francisco and Phoenix. Dolgov explains the technical evolution of Waymo’s driver, the safety-first pivot to full autonomy, the business model, and why he believes autonomous taxis will become a routine part of urban transportation.

Main Topics: Origins in DARPA and early autonomy research (Priority: 5/5): Dolgov traces Waymo’s roots to the 2004-2007 DARPA Grand and Urban Challenges, which forced teams to build vehicles that could navigate increasingly dynamic environments and seeded much of the modern self-driving ecosystem. Google’s secret self-driving project and early milestones (Priority: 5/5): He describes the early Google mandate: hit 100,000 autonomous miles and complete 10 100-mile routes without human intervention, which validated the concept but exposed many edge cases and failures. Why Waymo pivoted from driver assist to full autonomy (Priority: 5/5): The team moved away from assisted driving after seeing users overtrust the system and after gaining enough technical confidence to pursue the harder goal of fully autonomous driving in complex urban settings. Engineering the Waymo driver and safety stack (Priority: 5/5): Dolgov explains how cameras, radar, and lidar complement each other, and why Waymo chose multi-sensor fusion and purpose-built safety hardware to handle fog, rain, night driving, and complex city conditions. Commercial rollout and current operations (Priority: 4/5): Waymo’s technology evolved through multiple generations, leading to public autonomous ride-hailing in Chandler, Phoenix, and San Francisco via Waymo One, with fleets of roughly 250 vehicles in each city. Business model and future applications (Priority: 4/5): Waymo sees itself as building a generalizable driver that can power ride-hailing, trucking/deliveries, and eventually personally owned vehicles, rather than becoming a car manufacturer. Safety, liability, and the future of transportation (Priority: 5/5): Dolgov argues autonomous vehicles can reduce the status quo of roadway deaths, places liability on Waymo when its system is at fault, and envisions a future where most urban trips are handled by autonomous cars.

Key Arguments: Autonomous driving became feasible through iterative progress from controlled challenges to real-world city deployment, not through a single breakthrough. Waymo abandoned a limited driver-assist strategy because users overtrusted it and because full autonomy promised a more meaningful safety and transportation impact. Multi-sensor fusion (camera, radar, lidar) is superior for full autonomy because each sensor compensates for the others’ weaknesses in weather, light, and visibility. Waymo’s role is to build the driver software and system, not to become a traditional car company. Public ride-hailing deployments in San Francisco and Phoenix demonstrate that autonomous cars can operate safely and at scale in dense urban settings. The real transportation system is already unsafe, so autonomy should be judged by whether it materially reduces crashes and fatalities. Waymo accepts responsibility when its driver is at fault, and existing insurance frameworks can handle cases where another actor caused the crash. The long-term opportunity is broader than ride-hailing: trucking, delivery, and eventually personally owned autonomous vehicles.

Data Points: DARPA Grand Challenge distance: 150 miles - Early autonomous vehicle competition in a desert environment described as static compared with later challenges Google autonomy goal #1: 100,000 miles - First major mandate given to the Google self-driving team Google autonomy goal #2: 10 routes of about 100 miles each - Second mandate to complete devious, carefully selected routes with no human intervention Driver-assist pilot users: About 100 Google employees - Pilot program in 2011-2012 before Waymo pivoted to full autonomy First public autonomous ride: 2015 - Blind passenger Steve Mann took the first autonomous public ride in Austin, Texas Vehicle speed limit: 25 mph - Low-speed Firefly prototype used in early fully autonomous testing Waymo Safety driverless miles: Well north of 5 million miles - Distance Waymo says it has driven in fully autonomous mode Independent study miles: Almost 4 million miles - Miles reviewed by Swiss Re in its safety analysis of Waymo operation Bodily injury claims reduction: 100% reduction - Swiss Re finding for Waymo’s fully autonomous operation in the study period Property damage claims reduction: 4x reduction - Swiss Re finding for Waymo compared with human-driven benchmarks San Francisco fleet size: About 250 vehicles - Approximate fleet size in San Francisco, though not all are on the road at once Phoenix fleet size: About a couple hundred vehicles - Approximate fleet size in Phoenix U.S. annual traffic deaths: More than 40,000 per year - Dolgov’s argument for why the current transportation system is unacceptable U.S. personal vehicle idle time: 90% of lifetime - Used to argue that autonomous, shared vehicles could improve efficiency and urban space use U.S. miles traveled: Almost 3 trillion miles - Estimated travel volume in the United States that Waymo wants to impact

Pivotal Quotes: "I think I'm much more anxious about human drivers than I am about the Waymo driver." — Dmitry Dolgov: His response to concern about riding in an autonomous taxi "We are not okay." — Dmitry Dolgov: His argument that current roadway fatalities show the status quo is unacceptable "Our mission, which is to build the Waymo driver and then over time, deploy it in different products and different commercial applications." — Dmitry Dolgov: Waymo’s long-term business and product strategy

Implications: Waymo is positioning autonomous driving as a safer, scalable transportation layer rather than a novelty. If its safety claims hold, autonomous taxis, freight, and eventually personal vehicles could reshape urban mobility, liability, and how cities use road and parking space.