An interview with Waymo co-CEO Dmitri Dolgov on Tesla fans' claims that Waymo's software relies on hand-coded rules, can't handle freeway driving, and more
and soon tens of thousands — of vehicles as it expands to new cities.” https://www.understandingai.org/ ... Bluesky: David Hamilton / @davidneilhamilton.bsky.social : This article contains some great insights into the past technological/AI approaches used by Waymo & Tesla. — I'm not in 100% agreement with some of Timothy's conclusions (he glosses over the constraints that highly detailed mapping needs impose on Waymo) [embedded post] X: Josh Morrison / @joshcmorrison : @MattBruenig This Tim Lee piece was quite good - https://www.understandingai.org/ ... — the pro-Tesla case is that Waymo's cars are expensive and it will be harder for them to scale up manufacture than it is for Tesla to figure out software — but agree Waymo is underhyped and Tesla is ... not underhyped Jordan McGillis / @jordanmcgillis : Key point here from @binarybits: Tesla already has millions of vehicles ready to run its self-driving software once it's good enough; Waymo is trying to build a fleet from scratch, and that's going to be a slow and difficult process. https://www.understandingai.org/ ...
Context & Ripple Effects
Waymo and Tesla have long represented contrasting autonomy strategies: Waymo’s earlier position was that Tesla’s lidar-free approach would not yield a fully self-driving system, a divide captured in Waymo’s earlier rejection of Tesla’s approach as a path to full autonomy.
The interview lands in a sector where large investment has not reliably translated into commercial progress; the industry’s record of costly shortcuts and continuing losses makes technical claims about scalability and operating scope central to competitive credibility.
First-order effects
- Waymo gets a public forum to contest the characterization of its system as dependent on hand-coded rules and unable to operate on freeways, reinforcing its technical positioning as it pursues city expansion.
- Tesla and its supporters face a more explicit comparison of competing autonomy architectures, though the interview itself does not independently settle the performance claims.
Second-order effects
- The rivalry shifts attention from broad claims of “self-driving” toward demonstrable operating capabilities—such as the environments a service can handle and how it scales across cities.
- Potential riders, partners, and investors have stronger reason to scrutinize deployment evidence and operating constraints rather than treat sensor choices alone as a proxy for autonomy.
Third-order effects
- If commercial deployments increasingly become the decisive proof point, autonomous-vehicle competition may favor companies able to pair driving software with the operational infrastructure needed for repeatable city launches.
- The sector’s long-running split between camera-led and lidar-inclusive systems may matter less as a branding debate than as a test of which stack can sustain safe, scalable service operations.
The trend: Robotaxi competition is moving from architectural arguments toward evidence that an autonomy stack can operate and expand as a real service.