The future of computer-driven cars and deliverbots
The unusual thing about the Zoox is it is symmetrical -- it drives the same forwards and backwards. Now it's finally out on public roads. That's a good time to discuss whether it would be good for other, traditionally designed robotaxis to drive backwards for short stretches to get out of tight spots, to turn around, and to quickly get out when they discover a fire crew that will otherwise break their widows.
Driving is the hard problem. But doing pick-up and drop-off turns out to have a lot of complications and it was not at the top of the todo list, so some companies are having issues with it with cities. We see some hints of this in Waymo's Superbowl-related service, too.
Read more at Self driving cars have trouble with Pick-up/Drop-off, and for the Superbowl on Forbes.com
I have done an experimental podcast discussion show on the hot future-of-transportation issues so far this year.
You can watch it on YouTube, where I have chapter markers to let you easily find the topics of interest to you, as we spoke for almost one hour and 20 minutes. I was joined by Mario Herger of The Last Driver Licence Holder
On Jan 21, SF Fire Dept. crews, worried a Cruise robotaxi was about to drive through their fire scene, smashed in its window. They said it wasn't stopping, and back when Cruise first began one of its cars did drive over a fire hose. Digging into the details though, Cruise said it had stopped after trying to pull over, and did what they expected. So what should it do, and does the fact that that Cruise takes the conservative approach in such situations of stopping and waiting for rescue constitute a big safety problem, or just a teething pain as they test and learn.
An annoying paper argues that self-driving cars will use huge amounts of compute and thus have a giant carbon footprint. The boring way that it's wrong is that the compute load will not grow as they suggest.
The more interesting way that it's wrong is that self-driving EVs will draw most of their power from no-emission generation sources like solar and nuclear, even if they do use a lot of power.
I recently took a ride in a fully autonomous Waymo vehicle in San Francisco. It was my first ride in many years — I had been a member of the early team while it was part of Google. My guide on the ride was Andrew Chatham, whom I had worked with back then. He is now a Distinguished Engineer, managing fleet logistics and many other things, and reporting directly to Waymo’s co-CEO.
I have started building a map of all the autonomous services deployed carrying passengers or cargo. The services must be available to the public and out in public or semi-public spaces.
Turns out there are a lot. Contributions are welcome.
California recently passed a law that is obviously aimed at forcing Tesla to stop using the name “Full Self-Driving” to describe the expensive software add-on they sell for their cars which does not, at this time, provide self driving, full or otherwise. The ostensible reason for this is to avoid customer confusion and the potential danger that could come from people thinking they have a self-driving car when they don’t.
Many companies are working on self-driving trucks and delivery. As it became clear that robotaxi required a very large investment, teams looked for a more tractable problem. Most have gone after long-haul trucking on the interstate, but one leader has quietly gone after the “middle mile” while others go after long-haul and last-mile.
Cruise and Waymo, the clear #2 and #1 (respectively) in the Robotaxi race, have recently expanded their service areas for public access rides and driving with nobody in the vehicle. It’s a continued positive milestone in a year that has seen many setbacks for self-driving projects.
Here's a digest of some of my recent postings on Forbes.com
A filing suggests Tesla may be putting a radar back in their cars, but this time a high resolution radar, which is a bit like the LIDAR they swore was a crutch. It would be a good idea.
As a companion to yesterday's article about why the death of self-driving has been exaggerated here is an article asking what happens if the doomsayers are right, if people can't pull off a usable robcar and robotaxi for a decade more more.
There are lots of easier, more tractable opportunities out there, and I list a number of them.
This past month, especially with the shutdown of Argo.AI, have seen a number of declarations of the death of robocars. Thank to markets and expected consolidation, there definitely is a rough patch, but here's the argument that the field is hardly pining for the fjords and some things are going gangbusters, and not a decade or more away.
Read it on Forbes at Reports Of The Death Of Self-Driving Cars Are Greatly Exaggerated
Amazon's robotaxi company, Zoox, has always worked to be different, with its own custom vehicle designed from the ground up. They have added thermal cameras to it for night vision and detection of people and animals. I look at what that does and other factors about the normally low-profile company in this new Forbes.com article.
Tesla announced that new model 3 and Y vehicles delivered will no longer have the 12 ultrasonic sensors in the bumpers. They also disabled park assist and auto-park along with summon and the useless smart summon in these new cars, but promise those features will return soon as they work out how to do them with the cameras and software.
That's a remarkable move that no other auto OEM would do. Why have they done it and will it work? Read about it in a new Forbes site column at