Cruise complies with California DMV's request to halve its fleet, as the regulator probes “recent concerning incidents” involving its robotaxis in San Francisco
Cruise, the self-driving car subsidiary of GM, has been asked to reduce its robotaxi fleet by 50% in San Francisco following …
TechCrunchKirsten Korosec
Context & Ripple Effects
Cruise had been broadening service availability, including day-and-night driverless operations with a small San Francisco fleet. The DMV intervention turns fleet size into an immediate operating constraint rather than simply a launch-stage choice.
The restriction follows earlier calls from San Francisco for regulators to slow Cruise and Waymo’s expansion after vehicles stopped or idled in the city. It matters because local operational incidents are now directly shaping the pace of deployment.
First-order effects
Cruise must remove half of its San Francisco robotaxis from service while the California DMV examines the cited incidents, reducing its near-term operating capacity in the city.
The DMV gains a more active supervisory role over Cruise’s deployment, making continued scale contingent on addressing the regulator’s concerns.
Second-order effects
Waymo and other autonomous-vehicle operators face a clearer warning that service expansion can be curtailed when on-street performance creates city-level disruption.
Cruise’s rollout plans and the utilization of its existing fleet become less predictable, increasing the operational cost of proving reliability in a dense urban market.
Third-order effects
If regulators repeatedly tie fleet size to demonstrated safety and traffic performance, robotaxi deployment will be governed by incremental permissions rather than technology milestones alone.
This points to a regulatory moat: operators able to sustain public-road performance and regulator trust may gain an advantage, while smaller or less proven entrants face a slower route to scale.
The trend: Robotaxi commercialization is shifting from broad deployment ambitions toward regulator-controlled, city-by-city operational scaling.
DMV orders Cruise to reduce robotaxi fleet in SF by 50% after collision yesterday with fire truck, injuring passenger. Of course, just a handful of days ago the CPUC said Waymo and Cruise could vasty expand their fleets in SF. At least the DMV has some sense about this half-bak…
A little after 10PM last night one of our cars was in a collision at Polk and Turk St and we wanted to provide an update on what we know at this time. (1/5)
Last night a driverless vehicle collided with a fire truck in my district, sending a passenger to the hospital. This situation has gotten out of control. It's irresponsible & reckless to allow this experiment to play out on our streets without limitation
Saying “a driverless vehicle collided with a fire truck” is at best misleading and at worst a straight up intentional lie. The fire truck hit the Cruise car. I expect more honesty from a 5th grader. I know you won't respond btw, b/c you are also a coward.
to be fair who the fuck would get into one of those? they're playing crash dummies for their silicon valley overlords
Quote
·
Last night a driverless vehicle collided with a fire truck in my district, sending a passenger to the hospital. This situation has gotten out of control. …
Easy to posture when you can't do anything about it
Quote
·
Last night a driverless vehicle collided with a fire truck in my district, sending a passenger to the hospital. This situation has gotten out of control. It's irresponsible & reckless to allow this experiment to play out…
SF Chronicle publishes factually untrue headline The driverless Cruise car, which had a green light, was hit by an emergency vehicle. Yet the SF Chronicle headline claims the opposite. This is what you get when “objectivity” is abandoned by legacy media
If only they reduced the number of cars on the road by 50% after there was a crash involving a human driver. Cruise told by regulators to ‘immediately’ reduce robotaxi fleet 50% following crash | TechCrunch https://techcrunch.com/...
from @GaryMarcus “The degree to which you are safe in a driverless car still depends too much on the vagaries of data, and not enough on reasoning. (If that doesn't remind you of LLMs, you aren't paying attention.)” https://open.substack.com/...
Face it, self-driving cars still haven't earned their stripes. New post at Marcus on AI, on what has and has not changed over the last several years. [image]
We are (given new regulations) about to see a lot more edge cases pop up. It won't always be pretty. And, no, we won't probably succeed with deep learning alone, no matter how much training data.
An important thing to understand about car crashes is that they hardly ever have just one cause. Only one party might be at fault, legally speaking, but there are usually steps other parties could have taken to make a crash less likely. https://techcrunch.com/...
Here's a good example from last month. It sounds like the driver of the pickup truck would be legally at fault for rear-ending the Cruise. But if Cruise drove like a normal human driver (not stopping for shopping bags) this crash wouldn't have happened. https://www.dmv.ca.gov/...…