Vote 2020 graphic
Everything you need to know about and expect during
the most important election of our lifetimes

Apple’s Autonomous Car Apparently Only Works For Almost a Mile at a Time

Illustration for article titled Apple’s Autonomous Car Apparently Only Works For Almost a Mile at a Time
Illustration: Jason Torchinsky
The Morning ShiftAll your daily car news in one convenient place. Isn't your time more important?

GM and Apple are struggling to realize their autonomous dreams, Nissan and Renault are still figuring out their relationship, Volkswagen is slowly but surely rolling out its ride-sharing business, and more for The Morning Shift of Thursday, February 14, 2019.


1st Gear: Apple Seems to Be Having Trouble Doing This Self-Driving Thing

Most self-driving testing happens in the state of California, which also requires regular reports on that testing. One of the key metrics is miles between disengagements, or, the average amount of distance self-driving cars can go before their human drivers are forced to intervene, because the system is confused, or something else goes wrong. Waymo has far and away the highest average number of miles between disengagements, at 11,017 miles, according to the latest reports. Cruise is second, at a little less than half that.


What about Apple, whose Project Titan is supposed to also be developing a self-driving car?

From Bloomberg:

Test drivers disengaged the autonomous mode on Apple’s cars once almost every mile, based on data the company disclosed in an annual report to California’s Department of Motor Vehicles.

Once every mile! That’s astoundingly bad. If you have to take control of your “self-driving” car once every mile, that’s not called “self-driving” that’s just called “driving.” I could maybe see an engineering reason behind this—it’s possibly intentional, in other words—but in any light this is not where Apple wants to be. They also disagree with the metric:

Although access to data on the safety of test vehicles is key for the public to accept the technology, Apple told the DMV in April 2017 that its reporting requirements “do not achieve this result.” It suggested including metrics such as successfully prevented crashes. Companies that report to California also have made the argument that not all miles they test drive are comparable — highway driving on sunny freeways is much easier than navigating city traffic or snowy mountain passes.


This all depends really on how you want to define self-driving, which to me and any other human means a car that “drives itself,” meaning the driver should not have to do anything, whether it crashes or not. The disengagement metric accurately captures that, I think.

2nd Gear: Speaking of, GM Didn’t Reach Its Miles Target for Cruise

Cruise, GM’s self-driving unit, said in 2017 it was close to clocking 1 million self-driving miles per month. Fast forward to now, and, according to Bloomberg, Cruise has clocked less than half that figure all of last year.


From Bloomberg:

The mileage prediction was made 14 months ago by Kyle Vogt, who was then Cruise’s chief executive officer. In January, former GM President Dan Ammann took over as CEO and Vogt became chief technology officer.

In retrospect, the amount of driving Vogt described seems wildly ambitious. Even Alphabet Inc.’s Waymo unit, which also has autonomous vehicles on public roads in Arizona, only tested 1.2 million miles in California all of last year. While Cruise was founded in 2013, Waymo celebrates its 10th anniversary this year.

The good news for GM: Cruise has significantly increased the number of miles its vehicles are driving in autonomous mode without needing human intervention. The 5,205 miles logged between disengagements last year was second only to Waymo.

A spokesman for Cruise said the company is focusing on the quality of test miles and has been putting its cars through tougher driving conditions to improve its technology.


It seems like this self-driving thing is going to be a little harder to be, if it happens at all.

3rd Gear: Volkswagen Is Getting Into the Ride-Sharing Game

It is launching its service in Hamburg, Germany, in April. It’s called Moia, and will feature an electric bus with a 186-mile range. The bus will be driven by a human initially, but there are plans to make it autonomous (of course there are). There will be 100 buses to begin with, going up to 500 by the end of the year, with users ordering them through a smartphone app.


From Automotive News:

Germany’s second-largest metropolis after Berlin will serve as a test to see whether the business model functions properly and can be scaled up to expand to other cities.

“Ride-sharing services close the gap between taxis and public transit. Individual trips will be reduced and relieve urban traffic overall,” said Michael Westhagemann, Hamburg’s senator for economics, transport and innovation, in a statement.

VW has been operating trial runs of the service in Hamburg since last month.

Taxi services are as old as time but I’m still skeptical about the ride-sharing business model as it currently is constituted. I mean, call me when Uber’s profitable and we’ll talk.


4th Gear: Nissan and Renault Are Trying to Patch Things Up, Or At Least Renault Is

Renault’s chairman, Jean-Dominique Senard, is just now visiting Japan, the first such visit since the arrest of Carlos Ghosn in November. According to Reuters, Senard has told local media that he has no plans to try and take over Nissan as well, which is a little bit awkward since Nissan has a CEO, Hiroto Saikawa. But he was appointed before the Ghosn mess. And Nissan has no permanent chairman as of now.


From Reuters:

Senard said it was not the time to discuss whether he would also assume the chairmanship of Nissan, and that he expected discussions to be amicable, Jiji news agency reported, without giving direct quotes.

Senard was appointed chairman of the French automaker three weeks ago and is also expected to be named to Nissan’s board, given Renault’s 43 percent stake in the Japanese firm. He is expected to visit Nissan’s headquarters in Yokohama on Thursday and Friday to meet with board members and management teams.

The visit is aimed as a friendly, introductory call, according to sources familiar with the matter. But tension has been building between the two sides, with Renault initially backing Ghosn until he was forced to resign as chairman and CEO last month.

Some Nissan executives have long been unhappy with what they see as Renault’s outsized influence over Nissan - the Japanese automaker, in turn, holds a 15 percent, non-voting stake in Renault.


There’s no real there there with this story, so everyone’s still kind of in wait-and-see mode. I think it makes too much sense for the Nissan-Renault-Mitsubishi alliance to remain, but seems likely the terms might be different going forward.

5th Gear: Walmart Is Exploiting Its Drivers

The headline to this Reuters story is: “Why Walmart farms out same-day grocery deliveries to low-cost freelance drivers.” The answer is contained in the story:

[Jeff Leonard, a driver] and his cohort of some 16.5 million American “gig” workers - people who currently work in contingent jobs or as on call workers - come at a lower cost for Walmart than full-time employees, according to interviews with drivers, delivery companies and Walmart documents reviewed by Reuters.



How much do these drivers make, anyway?

“This affords me the ability to make my own hours,” said Leonard, who pays for his own fuel, car insurance and gets no health insurance, retirement plan or other employee benefits. He and other gig drivers in the area collect $7 to $10 per Walmart delivery. Walmart has deemed $11 an hour as minimum wage for its own employees.


Calling it the “gig” economy is marketing, and also the gig economy is garbage.

Reverse: A Star Is Born


Neutral: We Will Have Self-Driving Cars In ...

The year 2025? 2040? 2060? .... Never?

News Editor at Jalopnik. 2008 Honda Fit Sport.

Share This Story

Get our newsletter


Neutral :

I do not believe there will every be wide spread adoption of self driving cars. The personal expense involved is too high. The edge cases that are out of compliance require more sensing and more processing power, and will balloon the vehicle costs beyond the reach of the average, or even above average consumer.

There will be continued advances in ‘fancy cruise control’ (the systems currently available) but they will always carry the requirement that drivers stay engaged. Level 5 autonomy, if achieved, will require a price tag beyond anybody not shopping exotics.