Feds Question Tesla on Robotaxi Plan
Waymo “recalls” software for 1,200 vehicles, How Uber’s AV strategy could fail, and another Zoox crash
Hey, it’s Harry! Welcome to the 89 new subscribers who’ve joined The Driverless Digest since our last newsletter, pumped to have you along for the ride as we explore the intersection of autonomy and rideshare, and the business of AVs.
If you’re enjoying the newsletter, please share it with a friend. Let’s roll!
Top Stories of the Week
US agency asks Tesla to answer questions on Texas robotaxi plan (link, no paywall). The National Highway Traffic Safety Administration (NHTSA) has been investigating FSD collisions since last year and is now asking Tesla to explain how many vehicles it plans to deploy, how they’ll be monitored, and what safeguards are in place when visibility drops due to conditions such as sun glare, fog, airborne dust, rain, or snow. Tesla hasn’t commented yet, but this comes as scrutiny around FSD safety continues to grow — especially after a few fatal accidents involving the technology.
I think these questions are fair to ask of Tesla - after all, their vehicles will be operating on public roads so they need to be at least as safe as the average human driver. But safety, and even the definition of a robotaxi service are subjective. Instead, I’ll be looking at more binary questions for the June rollout in Austin, like who will own and insure these robotaxis and bear the risk and liability? If it’s Elon and Tesla, that tells me they have a lot of confidence in their latest FSD product.
A Zoox and an e-bike collided in S.F. Here’s what it says about robotaxi safety (link, no paywall). This incident happened in SF last month and only came to light recently. The crash appeared to be minor but it raises larger questions about robotaxi safety. Zoox shared footage of the incident with regulators but did not release footage to the public. And in a report filed with the DMV, they said:
A driverless taxi collided with a Bay Wheels e-bike whose rider had swerved into the car’s path. The Zoox sustained fender damage while the cyclist rode away. Nobody called police.
It sounds like the bike rider was at fault, although I would have liked to see Zoox release the footage to the public. To me, it matters who's at fault, since AVs shouldn't be expected to account for a biker running into a car. No human or AV could avoid that situation.
shared a different take:The question, as for any company, is how effectively Zoox will moderate its scaling to keep ahead of the problems they see on the road. Every incident is a chance to get unlucky. Those fixating on blame miss the point. Fatalities/injuries per mile have to do with net outcomes, not blame.
Cool Rides
Requested an Uber. They sent a Waymo. This is incredible. Drives perfectly. Better than a human. Didn’t know Uber and Waymo were linked (link).
This was a mind-blowing experience. Would you ride in a driverless car though? (link).
After spending some time in SF taking solely Waymo, I can’t even take regular Uber or Lyft anymore. The comparison of extreme danger when driving with those random people as opposed to the ultra safe self-driving cars is freaking me out. I can’t wait until the whole country has this option (link).
AVs/Humans behaving badly
This is how a Waymo reacts when it encounters a power outage at an intersection. I wonder how Tesla AI would handle this situation (link).
Cones no longer stop a Waymo (link).
Other Stuff
I discussed the possibility of this happening with Grayson Brulte last May and I think the analysis holds up well. Tesla doesn’t need Lyft but I think their user base and knowledge/operations/team could jump start Tesla’s rideshare efforts. As it stands, Lyft will always be in second place to Uber. But what if you could only hail a Tesla robotaxi on the Lyft network? That could make things more interesting..
Uber’s Vision for the Future of Autonomous Mobility and Delivery (link). My second TDD podcast is now live. Have a few more great guests lined up!
Waymo Issues ‘Recall’ On Robotaxis, But That’s The Wrong Word (link, no paywall). I’m with Brad on this one, recall is a terrible word for an OTA (over the air) software update. A lot of folks in the industry have come to ignore these recall notices because of this, but what happens when there’s an actual serious safety issue and a real recall notice needs to be issued?
How Uber’s AV Strategy Could Fail (link), by
. I like Uber’s AV approach so far. Most people already have Uber and Lyft on their phone, and they may add Waymo as a third option. And they may add Tesla as a fourth option because of the strong brand/leader. But will they add Zoox or any of the number of other companies working on AV tech? I don’t think so. A fragmented AV market benefits the major demand platforms like Uber and Lyft. The only problem is that Waymo is the only company in the United States doing paid driverless rides. So if Waymo decides to go at it alone, Uber is screwed.The road toward more driverless cars (link).
Uber is preparing its entire business for an autonomous driving future (link).
Aurora co-founder Sterling Anderson is leaving the self-driving truck startup (link) to become Chief Product Officer at General Motors (link).
With the Robotaxi network, your Tesla will be able to earn money while you’re not using it, essentially paying for itself. It will go to work, just like you (link). I’m torn on this one. Passengers already treat Uber cars terribly when there’s a human in it. I can only imagine what people would do if there’s no driver in the car (we’re already starting to see some issues with this on Waymo). Would you really want some drunk 20 year old puking in the back of your brand new Tesla?
The Waymo driver takes first trip to Boston and Dallas (link).
Shout-outs
Thank you
for featuring The Driverless Digest in Ride AI.Big thanks to TDD readers Daniel M and Charles W for referring new subscribers to TDD. If there's someone you think would enjoy TDD, just forward this email to them or use the referral button below.
Until next week.
-Harry