Waymo has applied for, and been granted, a licence to operate as a "Transportation Network Company" (fancy name for app-summoned taxi like Uber) in Arizona. This has been expected for some time, and shows they are continuing their plan to open up their pilot service in Phoenix to the public.

It's also (deliberately or not) as salvo against Uber (which they own even more equity in now) and even their new partner Lyft. Waymo/Alphabet have almost deliberately avoided getting into competition with those companies, even though of all the companies in the world, they are one of the best positioned. To provide a service that lets people say, "I will give up my car," you need to have a human-driven complement for the trips not on the robot network, and a seamless one way car rental service (like Car2Go/DriveNow) for longer trips. For now, this licence is probably there just to operate the Waymo robocars. But perhaps there is more?

No, they didn't run a red light

A video surfaced recently of a Waymo van making a mistake turning left at a red light. The short and boring answer is that that Waymo reports the car was being manually driven. When you want to turn left at a green, as you know you advance into the intersection but if you never get an opportunity to turn, you do it when the light turns yellow, and sometimes a bit into the red. This driver did a not very good job of this.

It's not impossible automated software might also make a mistake here, but the thing that makes it worth talking about is this. If the software had made a mistake like that, the safety driver would have logged it, and immediately the team would have met to develop a fix for it. Within days, all the cars would no longer make that mistake ever again. Think about how different that is from humans.

Highway deaths stay high.

Highway deaths went into a steady decline at the start of the century in the USA, but in the last 3 years, they have been climbing again. Recent estimates suggest 2017 will be about even with 2016, over 40,000.

The big decline was credited to a few things:

• Probably most of all, crumple zones, airbags and other injury reducing technology
• Doctors getting better at saving people
• Anti-lock brakes and stability control
• Last, and probably least, active safety systems like forward collision warning/avoidance etc.

I say least for the active safety systems because they just aren't very widely deployed yet. Recently, things like this became standard on popular new cars. When your car has these systems, they do make a serious reduction in accident incidence and severity, but most people don't have them.

The biggest culprit people blame for the rise is distracted driving. That's hard to measure of course, but there's not a lot out there to blame. In fact the only other thing to blame is scary -- namely people compensating for the safety systems by driving with less care. Of course we know about Tesla owners thinking they can look away from the road for a long time with AutoPilot. Collision warning and lanekeeping can make you more confident in writing a text.

Disengagement reports

This week I made reference to the Disengagement Reports from California. The reports once again show Waymo with a commanding lead -- and they don't even include all Waymo's new operations in Arizona.

The other companies are all hugely behind. The one exception is GM/Cruise, which shows a good climb and lots of miles being tested, but still way behind Waymo. Cruise will say that their numbers are even better because they do more testing in San Francisco's busy urban streets, and there may be something to that, but they still have a lot of distance to go.