Technology
The sensors on a Waymo are a real achievement. The business model wrapped around them is the problem. Build the technology. Build it around a driver.
Let's say it plainly, because the industry would prefer we didn't: the technology works. The Insurance Institute for Highway Safety, an independent group, looked at Waymo's data and found its vehicles are in crashes 68% less often than human drivers. Lidar sees through glare. Radar sees through rain. Software never gets tired at the end of a twelve-hour shift, and never checks a phone at a red light.
We want all of that. We want it in every car on the road. The question this site asks is who the technology is built to serve.
The same hardware that lets a robotaxi drive itself can sit in a human-driven car and watch the road alongside the driver: braking when the driver doesn't see the pedestrian, holding the lane when the driver's attention slips, flagging the cyclist in the blind spot, logging the near-miss so the fleet learns from it. Aviation has run this way for decades. Autopilot didn't remove pilots from cockpits; it made flying safer with the pilots still in them.
The industry chose the other path, and it was a choice. A system that assists a driver saves lives. A system that replaces a driver saves lives and eliminates a wage. Only one of those shows up as margin.
Artificial intelligence should extend what a person can do. The robotaxi is the version of the technology that removes the person instead.
Anyone who spends time on the streets of Los Angeles or San Francisco notices the same thing: Waymos driving with nobody in them. Some of that is expected. A robotaxi repositions between rides, heads back to a depot to charge, or waits for a fare. An empty car is not proof of an empty business. But the pattern raises a question the company has never answered in public: at any given hour, what share of its cars are carrying a paying passenger?
Documented Here is what we do know. Waymo's prices are not low. An analysis of ride offers in the Bay Area found Waymo averaging about 12.7% more than Uber, down from a 30 to 40% premium months earlier. Riders tell us the same thing: Waymo often costs more than a human driver. And Waymo's own case to investors is that more miles generate more data, which improves the AI, which enables expansion to more cities. Every mile has value to the company whether or not someone is in the back seat.
So the honest version of the question is this: is the robotaxi fleet on our streets today a rideshare service, or is it a data-collection operation that charges for rides when it can? If the second is even partly true, then the cars are being subsidized by investors to drive around learning our cities, and the drivers losing fares to them are competing with a product that doesn't have to make money yet.
We don't know the answer. We are going to find out.
In progress We are building a public log of robotaxi sightings. It takes thirty seconds and no app.
We will publish the counts by city, hour, and neighborhood, with the method, and we will send the results to the companies and to the California Public Utilities Commission with one request: publish your utilization data. If the cars are full, prove it.
Send sightings to: truthinmobility@proton.me
When a robotaxi company says its cars are safer, it is comparing its fleet to the average human driver, including the drunk one and the teenager. It is not comparing itself to a professional driver in a car with the same sensors watching the road. Nobody has run that test, because no company has an incentive to. We think it is the only comparison that matters, and we think the result would change the conversation.
Until then, accept the data for what it says: a well-engineered robotaxi crashes less than a tired commuter. That is worth something. It is not worth five million jobs.