GM’s Not-so-Super Cruise

“What’s he building in there?

What the hell is he building in there?”

-       Tom Waits, "Mule Variations"

Dan Ammann, CEO of GM’s Cruise operation, has been given the unenviable task of converting $7.25B and hundreds of employees (1,500+ by Ammann’s count) into autonomous vehicle industry leadership. It’s a sector that is seeing more disappointment than achievement of late.

In a new blog Ammann writes that the company has doubled down on mastering what Ammann asserts is “solving the most difficult version of the self-driving challenge” – operating in “one of the most complex urban environments” – San Francisco. (How about Mumbai?) He describes this environment as “40x more challenging than a simple suburban setting.”

https://tinyurl.com/yyrkr7jk - The Next Steps to Scale Start in San Francisco - Medium

The last comment (above) reads like a swipe at Waymo, which far surpasses Cruise’s performance based on miles driven between disengagements – the only available metric and one that has been bestowed on the industry by the California Department of Motor Vehicles. The difference between the two companies, ranked first and second by the CaDMV, reflects Waymo’s deployment in both suburban and urban settings – vs. Cruise’s urban focus.

Ammann claims the urban emphasis will enable Cruise to more rapidly deploy its vehicles “everywhere else.” That’s a bold and difficult to defend claim and it sets off alarm bells for any reader hoping to gain insight into automated driving from Ammann.

Ammann writes about the importance of trust and notes Cruise’s focus on “the building blocks necessary” to support its plans to rapidly scale its autonomous vehicle operation.  Ammann defines these building blocks as:

• Hiring “well over” a thousand of “the world’s best engineers;”
• Raising billions of dollars in capital;
• Achieving deep integration with General Motors;
• And focusing on testing and development in complex urban environments.

Those strategic building blocks are sound if one’s organization is targeting a well understood and commodified market or technology sector. The problem is that no organization has yet fully solved the autonomous vehicle challenge. Cruise will have to solve that problem first. 

Interestingly, Ammann chooses to not even address prior GM Cruise claims regarding the launch of a robotaxi service in San Francisco before the end of 2019. That prospect has been abandoned amid other missed milestones.

So there are at least three challenges facing GM Cruise which have nothing to do with Ammann’s “building blocks”:

• Solving automated driving in all use cases and environments;
• Understanding or defining appropriate use case scenarios and business models that will support and justify automated vehicles;
• Finding a way to compete with existing and lower cost urban transportation alternatives – taxis, buses, subways, scooters and bikes.

This is assuming, of course, that GM intends to get directly into the transportation business. Given the retreat of GM’s Maven car sharing service questions remain regarding the company’s ability to offer transportation services directly to the general public.

In his blog, Ammann goes on touting what he asserts is “already the world’s largest fleet of all-electric self-driving cars.” As of the end of 2018 GM Cruise reported 162 vehicles in operation to CaDMV. He adds that Cruise already owns nearly 40% of all EV fast chargers in San Francisco and is building the largest fast charger station.

The long and the short of the Ammann blog is that GM Cruise is building a massive fleet, a massive amount of infrastructure, hiring hundreds of engineers, and piling up self-driving car miles almost exclusively in a single urban setting. He further asserts deep integration with GM’s Warren Tech Center, claiming a seamless alliance with GM’s engineers working on mass produced vehicles thereby setting the stage to deliver self-driving vehicles on a massive scale.

The only thing missing is word one on how GM Cruise is developing and deploying its technology. A scan of the industry will find:

• GM’s Super Cruise (Level 2 automation) operation sharing details of its technology deployment and future plans at various industry events.
• Lyft (just yesterday) making freely available the “largest (automated vehicle) public data set of its kind,” “containing over 55,000 3D frames of captured footage hand-labeled by human reviewers, data collected by seven cameras and as many as three lidars depending on the car used, plus a drivable surface map and HD spatial semantic data that corresponds to the captured info to provide context to researchers.”
• Tesla Motors publicly describing its progress and plans in an hours-long session including engineers describing their work.
• Waymo and Argo have also shared data and insights regarding their work.
• BMW sharing safety data from its vehicles as part of the European Union Data Task Force effort.

It’s clear that GM Cruise could be doing much more to come clean with the industry regarding its progress and plans, rather than sharing meaningless marketing blather possibly intended to goose GM’s stock price. GM Cruise or GM itself could also be doing more to open up its data sharing activities to the industry in the interest of enhancing safety and saving lives. Surely there is an OnStar trove of data somewhere worthy of public scrutiny and industry interest.

Instead, the real takeaway from the Ammann blog is that GM will be continuing to further clog and congest San Francisco city streets with its growing fleet of self-driving vehicles which are doing nothing more than moving Cruise employees around while allowing the company to treat the city as its own private lab. Ammann claims to be performing a wide range of outreach activities. I think a deeper sharing of development progress and maybe even some data would go a long way toward achieving the level of public trust he claims is so important.

As far as operating autonomous vehicles in San Francisco or any other city is concerned, it is probably the one use case scenario that is the most difficult to justify given the plethora of lower-cost alternatives. It is also unclear as to how GM’s Warren Tech Center engineers plan to merge the Super Cruise Level 2-type solution with Cruise’s Level 5 technology.

“Delivering self-driving cars at scale isn’t just about winning the tech race, it’s about winning the tech race and the trust race,” concludes Ammann. Given GM’s recent history, winning the “trust race” may be an even steeper task than unravelling automated operation in San Francisco.