Alt title: The Data on Self-Driving Cars Is Clear. We Have to Change Course.

Gift link: https://www.nytimes.com/2025/12/02/opinion/self-driving-cars...

That would be great if we actually had self-driving cars. What we've got now is a mess, using the public roads for a development environment.

That’s very uncharitable. There’s a lot of in-lab development and outdoor (but not public road) testing.

Ultimately, however, you’ve go to insert into the real world.

Do you have a better idea?

Yeah, don't grease politicians' palms, and do a much slower, well-controlled rollout. Driving involves so many variables and edge cases that if self-driving screwed up as much as AI hallucinates, we'd all be in trouble. It's not just accidents, it's stopping in the middle of the road, disrupting traffic, annoying other drivers to the point where they're likely to do something dangerous to get around them. Instead of asking the patrons if they liked their ride, try asking local residents who have to share the road with them, or first responders, or pedestrians -- or pet owners.

What are the concrete statistics you have in mind that signal that we've been too aggressive in rolling out, cutting too many corners?

A slower rollout wouldn't help anything because the cars would learn slower.