Will Cowboys Or Collaborators Shape The Self-Driving Future?

You may have noticed the bloom is ever so slightly off the autonomous vehicle rose. This is likely due to some combination of a generalized malaise and growing cynicism directed at Silicon Valley in general (where of course much work on AVs continues apace) and a growing list of highly publicized self-driving incidents that surely make PR teams groan. Exhibit A from this week was the California Highway Patrol using Autopilot to pull over a dozing, blitzed driver — showing the universe has a sense of humor, or at least irony, this happened in the heart of Silicon Valley. Kudos to the officers for knowing how to communicate with the Tesla’s autonomous system when the driver proved oblivious to the flashing lights and siren, not a use case featured in promotional materials, I’m guessing.

It was around 3:30 in the morning when a California Highway Patrol unit pulled alongside and observed the inexplicable: The driver was asleep.

It was then that the officers invented what the CHP says is a new maneuver in highway traffic control. https://t.co/Gdhe71J2Fn

— Los Angeles Times (@latimes) December 2, 2018

Upfront here I must state that I’m with Elon Musk that eventually some sort of general-purpose, autonomous vehicle will arrive and change the world. Trying to produce such a vehicle is a worthy end goal. But I might respectfully take issue with his iconoclastic notion that technologists like him will have to march toward this goal alone. In his interview last month with Kara Swisher, Elon was asked about the prospects for infrastructure and road sensors being incorporated into self-driving solutions. His answer: “Yeah, that’s hopeless. That would, at best, be a specialized solution, and whatever city puts stuff in roads … You can always make something work for a specific solution, like some special-case solution in some town, you can make that easy, but what you really want is a general solution for self-driving that works worldwide.”

It’s a provocative comment, the kind that makes Elon such a compelling figure for the media, and that reliably lights up the Twitterverse. It also runs counter to the current moment, in which autonomous vehicles are very much paying their dues in the trough of disillusionment. And, perhaps surprisingly for someone involved in such protean endeavors as Elon, it reflects a fairly narrow view of the technology. Here’s why.

Considering an autonomous vehicle in isolation is like thinking about the engine of a car today without the rest of the vehicle. Eran Shir, CEO and co-founder of Tel Aviv-based Nexar, is among those betting that smart cities will be linchpins of autonomy. Earlier this year he told Forbes contributor Omri Barzilay “the cities of the future are going to require intricate and complicated management systems,” in an article on Nexar closing $30 million in series B funding.

The world’s megafirms, including Siemens, have also taken note of the vast potential of public-private partnerships. An October article in our own ‘Pictures of the Future’ magazine sketches out the Siemens vision, noting “increasing traffic volume is creating problems in cities around the world, congestion, poor air quality and less productive economies. Connected autonomous vehicles with electric drives could offer a solution…”

But let’s step back a little. I can already hear the Tesla faithful and Valley tech stalwarts asking: “what is the motivation for carmakers to get involved in cities when these firms already have their plates full with autonomous, connected and electrification R&D spending?” Let me answer with a market-inflected twist on Elon’s famous guidelines on thinking in a first-principles way. Looking decades out, the main objective any private firm in the transportation sector needs to worry about is how to maintain and grow their revenues. That’s about as far as we can abstract it.

Despite inevitable business, political and even environmental change, firms today and tomorrow need to make money for stockholders and investors as their raison d’etre. Period. And the best collective thinking about those future revenue streams suggests they will increasingly spring not just from vehicles (autonomous or otherwise), but rather from connected, autonomous, electric vehicles of all sizes and form factors operating within a smart city environment. The first-principle of a healthy balance sheet is to grow the top and bottom lines, no matter whether that means selling trips, experiences or physical cars. What will matter is having a differentiated product, which probably means a combination of physical cars, ride-sharing, trip-selling, last-mile delivery vehicles and who knows what other economic opportunities that we haven’t even thought of.

The good news for carmakers, from Tesla, to the Chinese upstarts to the old guard in Detroit and the word’s car capitals, is that vehicles remain at the center of this logic diagram — the nuclei around which swirl countless moving parts, from chip to city, to borrow an image that might be more appropriate in a discussion involving physics-loving Elon. In an interview for a recent Wall Street Journal article in the by-now familiar genre of will-Detroit-be-left-behind, Sam Abuelsamid, a senior analyst with Navigant Research, said “ultimately, you can have the best services platform there is, but if you don’t have the vehicles to operate on it, that won’t do you much good. That’s where the manufacturers have an ace in the hole.”

Sam, who has graciously shared some of his photos with our writers for various whitepapers, is right. However, carmakers need to actively play that ace as they grapple to stay central as the world collectively moves toward Level 4 and Level 5 autonomous vehicles. Once we get to full Level 5, with all road vehicles operating autonomously under all road conditions, we will have solved several challenges — how to handle that messy time window with a mix of autonomous and human drivers, how vehicle infrastructure and vehicle-to-vehicle communications operates flawlessly, how cars integrate cleanly with other transportation modes to provide frictionless trip planning, how autonomous vehicles and pedestrians live together safely on our city streets. And with these solutions will come a new normal for the industry, which likely will be marked by a small handful of entrenched firms with huge inbuilt network-effect and platform advantages, if this current internet and smartphone era is any guide.

So imagine that autonomous vehicles are a building block (albeit an important one), and not the end objective in themselves. For carmakers and their supply chain, this means a whole other set of parallel challenges to those outlined above. As has proved true with smartphones, starting at the micro level, firms must develop, protect and fully utilize IP that will be the foundation of their differentiation. Methodologies like high-level synthesis are powering the democratization of chip design for differentiated IP such as hardware accelerators for machine learning algorithms. And we see carmakers already investing in this area, understanding that the integrated circuits that power the sense/compute/actuate functions in future vehicles will be a key advantage for product differentiation and time to market. (Elon stirred up loads of interest in Tesla’s forthcoming AI chip, expected next year, according to The Verge.)

From the integrated-circuit level on up come new verification, validation and integration challenges related to vehicles’ embedded systems requiring simulation tools and generative design methodologies for safe and secure operation. This control-your-destiny-future also means that the vehicle behavior will need to be modeled and proven robust long before taking the prototype to the testing and proving ground. It means that vehicle integration into the smart city environment (including traffic modeling, fleet management and vehicle performance data), will have to be managed in a cloud-based system, feeding data back to the design and manufacture sites for product optimization. And talking about manufacturing, how do we move from today’s large-scale car production, towards small batch, reconfigurable, customizable vehicle interiors, optimizing materials flow, time and comprehending new manufacturing techniques that will become mainstream, such as additive? Tony Hemmelgarn, Wally Rhines and Marcus Welz had an excellent chip-to-city discussion on these points at our IESF conference in Michigan a few months’ back; it’s here if you missed it:

There are two points to all this: firstly, the autonomous vehicles do not exist in a bubble. They are part of a wider transportation future that has many parts in play right now. Sure, autonomy (rightly) gets a lot of the tech media attention right now as it’s cool, futuristic and relatable, but we need to take a broad view when looking at how it plays into the future of mobility. Despite his go-it-alone stance, I suspect that Elon, whose transportation endeavors go from subterranean to space, knows this better than most. Secondly, solving the basket of challenges impacting all of us will require technology companies to take a broad perspective, and ideally invest in all these areas, while also playing an active part influencing standards bodies (e.g. safety and security), legislature, cities. The winner will likely be the one that pulls together all parts of the supply chain, from the complex integrated circuit manufacturers, through the tier 1 suppliers and the carmakers themselves, enabling IP sharing and co-development.

Kara Swisher has yet to turn her gaze this way. But if she ever does, expect someone from Siemens to emphasize the future is about keeping the broad perspective, and focusing on a deep understanding of how often conflicting elements can come together, creating value through a collaboration between old systems (like human-powered CHP cruisers) and new ones (Teslas with ever more powerful Autopilot) alike. Expect this future to be daunting, lifesaving and more than occasionally maddening — and to continue to shape up as the opportunity of a generation.