An unprovable claim: 95 percent of the existing literature on driverless cars is written in the future tense. Soon cars will do X, passengers will do Y, and so on. For some, I imagine this stylistic tic generates excitement—look at all the cool stuff that will happen! As a tech skeptic with a nine-year-old laptop, I can’t help reading chapters full of future tense and hearing a robotic, monotonous tone: “We will be with you shortly.” The future tense lends itself to these tonal interpretations, usually because the focus is on possibilities instead of outcomes. Reading autonomous vehicle literature is like watching clouds: You see a unicorn, I see a tank; neither of us is wrong, necessarily.
Except, why are we out in some field looking at clouds? Don’t we have work? Class? Kids? According to the existing autonomous vehicle (AV) literature, the answer is: What are those? A quick survey of the field reveals less about the underlying technology of AV and more about how the AV industry views the available time and resources of the average consumer.
Take 2016’s Driverless: Intelligent Cars and the Road Ahead, which offers several hundred “exciting” futuristic scenarios, almost all of which I find terrifying. For instance: Authors Hod Lipson and Melba Kurman theorize that the “crude, poorly designed in-car infotainment systems of today’s vehicles will no longer exist,” and whereas I envisioned the total eradication of this strange, possibly meaningless thing called “infotainment,” the authors quickly assured me that in driverless vehicles, the infotainment systems will not be eliminated but improved upon. Later, Lipson and Kurman envision a driverless taxi that will feature “a 200-page click-through license” wherein the passenger waives the right to sue; to me, this sounded like a dystopian Disney ride for lawyers, yet the tone in their future-focused paragraph is oddly rosy.
These are petty tonal disagreements compared to a scenario that unfolds late in Driverless. In what might be the most telling line in all of AV literature, the authors outline opportunities for in-car advertisements. Lipson and Kurman posit that by the passengers logging into the driverless vehicle system, the car will know all sorts of consumer information about them. The car will beam this information out to businesses appearing along the route to the destination. Driverless then concocts a can’t-miss deal for the passengers: “Restaurants will agree to split the cost of gas for a family’s road trip if their digital profile indicates that each time this family dines out, they spend $200.”
What’s telling isn’t the scenario. It’s the “$200,” which broke my readerly rhythm like a seat belt in a car door. Your city and acquaintances may vary, but I do not know a single family that spends anywhere close to “$200” “each time” at dinner. I started wondering: Do they mean a family of eight? Did I miss a reference to some kind of futuristic inflation?
They didn’t, and I didn’t. They wrote what they meant, i.e., $200 dinners as a baseline expectation for the average family. The line is a perfect, almost comic insight into the classist world of AV writing and AV thinking, where an embarrassing number of arguments do not seriously engage with how the coming AV revolution will affect the middle class, let alone the lower ones. In some instances—and I am only 99 percent kidding—one wonders if Driverless is actually written for cars. On multiple occasions, the verb “liberate” is used. As in: “Cars will be liberated from human drivers.” If the official tense of AV literature is the future, then the tacit audience may be the vehicles themselves.
There is one fleeting, acceptable use of “liberated” in Lawrence D. Burns and Christopher Shulgan’s Autonomy: The Quest to Build The Driverless Car—And How It Will Shape the World. It takes place during a fraught moment for Carnegie Mellon University’s Red Team, a group of engineers and roboticists scrambling to prep a driverless vehicle for the first DARPA Grand Challenge in March of 2004. Their robot, “Sandstorm,” has been smashing into poles, catching fire: “Sometime later, with Sandstorm liberated from the barbed wire and the deadline approaching, [Red] Whitaker gathered Urmson and everyone else around him.”
There’s nothing extraordinary about that line. But it is scenic, and in the field of AV lit, this is extremely rare. Autonomy’s ability to balance hypotheses with scenes ranks high on the list of reasons that it is far and away the best of an admittedly so-so group of AV books. It spends at least half of its pages covering every snag and success in the 2004, 2005, and 2007 DARPA Grand Challenges. There are “Getting the Band (of Nerds) Together” scenes. There are “Getting the Band (of New and Other Nerds) Back Together” scenes. There are races written tensely and races that are campy (squads of highly accomplished engineers skulk nervously for hours in the desert because none of the event organizers had the extremely lo-tech foresight to provide video footage of the course). After the races, when the respective teams disband to more or less originate the AV industry, there is workplace drama.
In short, Autonomy delays its “will” for a long time. It’s rooted firmly in “was.” Yes, there are some technical sections, but they end, and when they do, they often conclude with cliffhangers: “Thrun had no idea, at that point, that Page would change the course of his life.”
A bit cheesy? Sure. But it’s purposeful cheese. For a text with intermittent chunks of technicality and calculations, the book has an undeniable velocity, in part because Shulgan, the reporter with a ghostwriting credit, knows how and when to ramp up the narrative stakes.
It’s also probably rather important to note that Autonomy radically altered my perspective on AVs, largely because of Lawrence D. Burns. For 30 years, he worked in General Motor’s research and development department, yet he refuses to cast himself as a “car guy.” He states clearly and repeatedly that what he’s interested in is mobility, which is to say that he has privileged humans over machines, both in his career and in Autonomy. He makes a number of irrefutable arguments about the environmental, safety, and financial benefits of AVs, but I have to confess that I was most shaken by his framing of the profit-motivated “car guys.” When Burns describes the incredible resistance to change on the part of Detroit car guys for much of the previous decade, I thought, “That sort of sounds like me.” I experienced a sinking feeling: If I am stubbornly against AVs for reasons that amount to a line in a book about “$200,” am I siding with … the car guys? Them? Over there? That’s me?
So Autonomy is good. It’s the best AV book to date. It is scenic and reported like the best pop nonfiction. It laces an evidence-based polemic throughout the DARPA narrative. It provides an insider’s account of both Detroit’s “car world” and Silicon Valley’s “mobility world.” And—if I can be allowed a bit of future tenseness—I think the book will prompt two interesting discussions, both of which have to do with sources.
The first source-related issue will, in true American fashion, be both a bigger deal to many people and also much less relevant. It concerns biases on the part of Burns. While at General Motors, Burns had business relationships with many of the technologists interviewed in Autonomy. After leaving GM, he consulted for Google’s Chauffer project, which later become Waymo, Google’s self-driving technology development company. It’s clear he personally knows over half (if not all) of the people in this book. Whereas Burns, a Midwesterner, frequently refers to Canadian wonderboy Chris Urmson as a “straight shooter,” other engineers clearly irk him. Chris Levandoski’s leadership at Uber is said to be “dominated by mistakes and accidents.” Tesla and Elon Musk receive similar criticism. It seems likely that some in the AV industry will think favorites are being played.
Which is a silly criticism. Thanks to publicly viewable lawsuits, it’s pretty clear that when it comes to individuals like Levandoski—who was sued for stealing 14,000 technical files as part of his transition from Waymo to Uber—Burns is actually quite large-hearted. Conduct a news search for “Tesla autopilot crash” and you’ll likely find results that are both more damning and more current than anything in Autonomy.
But to seriously interrogate Burns’s biases is to miss the point. We should want veterans of innovation with specific points of view turning to book-length work. Moreover, we’d be lucky if all such efforts are as structurally nimble as Autonomy. When Burns enters the text a fifth of the way through—“So in 2006, Whittaker came and visited me at the General Motors Technical Center”—the “me” is exciting. Suddenly, the book becomes a memoir. Then it slides back into reportage, then back to an insider’s tell-all, and there’s also math, plus small essays about labor, and then urban design, and also hearing aids, and so on. Whereas industry types might see conflicts of interest, readers should see a thrilling interaction of genres within a single title. Burns throws quite literally as many narrative styles as he can manage at an issue he feels is era-defining. Though tech sites do admirable work reporting on the ills of the tech industry writ large, their highly curated “vehicle experience” tours seem silly when compared to book-length work that has room to examine its subject along multiple lines of inquiry.
Autonomy does have at least one predictable oversight, though. It will surprise absolutely no one that the second source-related issue in a tech-sector book concerns women, specifically the scant attention they receive in the narrative. The acknowledgements page is a slew of men. Women are quoted maybe once every 60 pages in Autonomy (this ratio is almost surely generous).
Individually, the lack of quotes is at times justifiable. Collectively, the effect is predictably icky. The labor of Chris Urmson’s wife is rendered in a parenthetical: “(He and Jennifer had since had a second boy).” Later, when a secretary cries upon hearing about a setback in advance of another “big race,” I thought, “Gosh, it would be great to hear her perspective on why she cried, how much she devoted to the project, the hours she worked.” One of the engineering teams features a married couple, and I thought, “What is it like to be the lone woman on a history-making robotics team and to work alongside your husband and for both of you to be juggling these insane work demands and also various lucrative job offers?” Where is that story?
Again, in every scenario, there are no doubt explanations for why the source was not consulted. It’s possible that interviews were attempted and denied.
But if we reach for justification—“Well, the secretary was not on-site for many events”—then we know the exact type of book we will be reaching for on the new-releases shelf. It will feature men. It will not be misogynist, necessarily, but like many successful if limited initial textual offerings about a phenomenon, it will have a particular lens. It’s easy to mock the tech industry’s reluctance to champion minority perspectives, but the life cycle of book-length coverage is equally slow to highlight the missing voices. Whether it’s cars or the internet or outer space, the cycle for innovation-centric texts is familiar: Here’s what “will” happen. Then, a definitive white, male-centric tome emerges. Ensuing books focus on the male figures treated unfairly in the tome, then, years if not decades later, the cracks get filled in, the marginalized voices heard, just in time to be overshadowed by the emergence of a new innovation. In the future, a landmark book in an emerging tech-related field will hopefully be confident enough to skip ahead, i.e., a writer will draft the tome, identify the cracks, then discard the tome and essentially write what is missing. If this laborious process sounds like overkill, consider that the alternative is to lay a foundation with built-in cracks.