11 C
New York
Tuesday, February 27, 2024

The Relentless Startup Fast-Tracking Ford’s Self-Driving Cars

Like the residents of many cities that boomed before the advent of cars—and before the arrival of the coronavirus—the people of Pittsburgh get around using a jumbled network of too-narrow two-way streets laid out across a hill-studded landscape. It’s the kind of mess that would put Baron Haussmann in a mood for demolition, but the driver of the white Pontiac sedan coming toward the car I’m riding in reacts with generosity, slowing down and flashing his lights. His offer—ceding his right of way to let my car turn left ahead of him—is a kind one that is entirely incomprehensible to my driver.

I’m in a white Ford Fusion, the sort with the rooftop ziggurat of lidar laser scanners and cameras that identifies it as a self-driving car. And while those sensors could detect the Pontiac’s lights just fine, the software that turns their field reports into battle plans isn’t designed to consider this oh-so-human communication. It’s focused on the fact that the Pontiac is still moving forward, albeit at a crawl. And it’s not about to put itself in front of an oncoming vehicle. So we wait. After a moment, the Pontiac driver decides we’re a lost cause and accelerates past.

I’m sitting in the back seat with Brett Browning, the Australian native who is the head of robotics at Argo AI, the outfit that programmed the Fusion to drive itself. “The AV is still pretty conservative,” he says. With safety operator Melinda behind the wheel, ready to take control if necessary, the car’s AI finally decides it’s safe to turn left. Conservatism is a logical approach for a nearly 2-ton robot exploring Pittsburgh, but Browning can’t settle for safe.

He needs a machine that can offer an efficient, smooth ride to paying customers—whenever Argo is ready for those. “You do need to be assertive,” he says. “You need to drive confidently.” So the car waits until relatively late to hit the brakes before a stop sign. Every time a light goes from red to green, it zips up to its cruising speed, perennially set just below the legal limit.

The riders who will someday replace Browning and me in the backseat may never hear the name Argo. While competitors like Waymo, Uber, and Cruise intend to operate their own robotic ride-hail services, Argo is focused on making the car drive itself. The logistical and marketing work required to roll out a service and run that kind of fleet will fall to Argo’s chief customer, the Ford Motor Company.

The Detroit giant is more than a customer. On a Friday morning in February 2017, Ford’s then-CEO, Mark Fields, announced that the automaker was investing $1 billion, over five years, in Argo. The dollar figure and terms of the deal made clear that Fields and his executive suite didn’t just consider self-driving key to Ford’s future. They had accepted that they didn’t have the capability to develop it themselves. Facing a fast-shifting landscape, Ford needed Argo to deliver and was willing to pay handsomely.

What made the news odd was that at the time, nobody knew Argo existed. The startup had no cars, little software, and few employees. It had been known as Argo and had a proper office for only a few days. (Its original name, Spin Mobility, was just a placeholder.) But anyone who knew the self-driving field could understand Ford’s thinking. The automaker was looking not at Argo’s present but rather at its potential. Potential for which it was willing to spend the kind of money that could produce a new line of pickup trucks or SUVs. Potential almost entirely contained in the brain of a soft-spoken then-37-year-old engineer named Bryan Salesky. (He’ll turn 40 next month.)

I took my ride last autumn: Like its competitors and the rest of the country’s so-called nonessential businesses, Argo has gone into a sort of lockdown amid the Covid-19 pandemic. The company has suspended testing on public streets and curtailed work at its private test track outside Pittsburgh. Its software engineers are keeping progress going from their homes, running more than 1 million miles in computer simulation every night. The hardware folks are figuring out how to do their work outside the lab; one engineer has set up a moving target system in his backyard to evaluate a radar system. But a temporary slowdown shouldn’t overly hinder an effort with the kind of support necessary to playing the long game of making a self-driving car.

Given his early career trajectory, you might be surprised that Bryan Salesky now finds himself at the forefront of the race to deliver the self-driving car. Red-haired and blue-eyed, he was born in the Detroit suburb of Woodhaven, Michigan, where his father did factory work in a steel mill. As the steel industry cratered in the 1990s, his mother remarried and the family moved around, winding up in Pittsburgh.

Salesky earned a Bachelor of Science in Engineering at the University of Pittsburgh in 2002 and, uninterested in continuing the kind of education that meant writing papers for professors, decided against grad school. Instead, he took a job at Union Switch & Signal, the company George Westinghouse founded in 1881. There, Salesky worked on software that kept trains from colliding while traversing “dark territory,” the long stretches of track ungoverned by signal systems.

In 2004, a friend suggested he check out the National Robotics Engineering Center, an arm of Carnegie Mellon’s Robotics Institute. Salesky was skeptical—he wasn’t interested in doing research at a university—until he went in for an interview. NREC (pronounced en-wreck) made robots for commercial uses, catering to clients such as the US Army and John Deere. One project was Spinner, a remote-controlled, six-wheel tanklike creation that plowed through ditches, mowed down small trees, and could keep going if turned upside down. “I was 24 at the time,” Salesky says. “I’m like, this is what I want to do with my life.”

Equally impressed was Peter Rander, the NREC exec who interviewed Salesky for the job and who now serves as Argo’s cofounder and president. Rander found that the young engineer’s résumé was short, but his knowledge of the Union Switch & Signal railroad system was striking in its breadth and depth. When Rander asked how he had tested his work before launching it—and taking lives into his hands—Salesky described an extreme form of pressure testing for an easy-to-ignore problem that, given enough time, cables might fray or degrade. He sent a colleague into the server racks to start pulling out network cables at random. “That was some of the most fun stuff,” Salesky told Rander, who soon became his new boss.


Salesky first stepped into the world of self-driving cars in 2006, when an NREC leader told him to meet Chris Urmson for a cup of coffee. Urmson, a bit older than Salesky, had been a leader on Carnegie Mellon University’s efforts in the 2004 and 2005 Darpa Grand Challenges, the races that sent autonomous vehicles rollicking (and crashing) through the Mojave Desert. Now he was the technical lead on CMU’s bid to win the 2007 Urban Challenge, which ratcheted up the difficulty by putting the vehicles in a mock city, replete with intersections and parking lots and populated by human drivers, as well as other team’s robots.

For the 2004 and 2005 efforts, Urmson had written much of the software himself. To tackle this more complex challenge, CMU stocked the team with more than a dozen senior researchers, naturally independent types who had a bias for exploring novel techniques. Urmson, as a newly minted PhD, had little experience keeping such a group in sync and on task. That posed a potential problem. Unlike any research project, the team couldn’t settle for a one-time demonstration that some new idea was valid. And they couldn’t extend the deadline. Winning the Urban Challenge meant producing a robot whose reliability rivaled that of a commercial product, one guaranteed to work as intended on the day of the competition, for 60 miles and six hours. Administration higher-ups, frustrated by losses in the 2004 and 2005 Challenges, put Salesky in charge of software.

Salesky was slotted in as Urmson’s lieutenant. The man who had kept real trains on their tracks would now keep the metaphorical ones on schedule.

One of several team members whose formal education had stopped with a bachelor’s degree, Salesky thought more like a product manager than a researcher. Results outranked ideas. He brought in tools that would let the 20 or so members of his software team write code in parallel, then integrate and test their work. He made them create detailed plans for their projects, then saw that they stuck to them. “He really brought a professionalism to the Urban Challenge,” says Kevin Peterson, who did much of that software work. “Before that, we were kind of running at things as fast as we could.”

Salesky’s methodical attitude brought him into regular conflict with Urmson. Salesky was impressed by Urmson’s work but knew the researcher had minimal experience making products, where reliability mattered as much as capability. Once, he came in to find that Urmson had spent a weekend reconfiguring how the robot processed data from its lidar scanner. His reaction was to worry over what could go wrong: How the new code might break down or falter, what other parts of the system it might inadvertently compromise, what other software had to be updated in some way to work with it.

It was the sort of thing that led to frequent arguments between the two. “I wanted to make sure he wasn’t pulling in so much science and theory, to make sure that at the end of the day we were going to get something that was robust, that would satisfy the mission,” Salesky says. But he knew beating Stanford, MIT, and other top-tier teams would require inventive thinking. “Had he not done that, we probably wouldn’t have had the right perception system.” As CMU’s robot—a Chevy Tahoe festooned with sensors and named Boss—came together, the two learned to appreciate each other’s point of view and to balance their competing drives. They became close friends and a formidable duo. Their teammates dubbed them “momma bear (Urmson)” and “papa bear” (Salesky).

When race day rolled around in November 2007, Boss was set to leave the starting gate first, after an impressive showing in the qualifying round. Salesky was in charge of launching the software, and, always obsessed with quality control and detail, he checked that everything was running properly multiple times. A few minutes before the flag dropped, before the robot entered the field and escaped its creators’ control, he decided to look one more time.

To his horror, Salesky realized that the GPS had crashed, stripping the vehicle of its ability to navigate. He, Urmson, and others scrambled, replacing chunks of hardware and running diagnostic tests, while rumors of what was wrong snaked through the grandstands. Finally, the group looked beyond the robot and realized that the Jumbotron that Darpa had set up for fans might be interfering with the GPS radio signal. They asked race officials to kill the screen, and the robot’s GPS sparked back to life. Six hours later, the car crossed the finish line, giving Salesky, Urmson, and the rest of the Carnegie Mellon team the victory over rivals like Stanford and MIT, along with a $2 million prize.

Salesky walked away from the win amazed by how much the team had accomplished in 18 months and thinking about what they could pull off next. “The problem,” he says, “was funding.” Darpa was done with its Challenges. General Motors, which had sponsored Carnegie Mellon’s effort, declined to pay for a bigger push on autonomous driving—at the time, it was hemorrhaging money and headed for bankruptcy. Salesky and Urmson ended up launching a project for Caterpillar, developing autonomous trucks for use at mining operations.

They were just getting that project up and running when they received an invite to the Lake Tahoe chalet of Sebastian Thrun, who had led victorious Stanford’s teams in the Darpa Challenges. Thrun, who brought along his most impressive teammates, told the half dozen men assembled there that Google cofounder Larry Page wanted a self-driving car and was ready to pay whatever it took to get one. The Stanford group, some of whom were already working in some capacity for Google, signed on. So did Urmson.

Salesky demurred. He wasn’t ready to uproot himself and move to California for what didn’t seem a sure thing. Plus, he worried that if he and Urmson both left, the Caterpillar program would founder. It wasn’t until more than two years later, in 2012, that Urmson convinced Salesky to come out west. By that point Caterpillar’s autonomous mining trucks were en route to commercialization, and the engineer was ready for a move.

When he did join the Googlers on what was then known as Project Chauffeur (in 2016 it became Waymo, a standalone company under the Alphabet umbrella), Salesky stayed just a year. He soon tired of the tension between Urmson and Anthony Levandowski, who were fighting for control of the team after Thrun shifted to other work. (Levandowski’s 2016 move to Uber would spark a bruising legal fight between Waymo and the ride-hailing giant. The parties settled in February 2018, and in March this year Levandowski pleaded guilty to one criminal charge of trade secret theft.) Salesky retreated east, only to rejoin Google about a year later, this time to supplant Levandowski as hardware lead—Urmson had won the power struggle.

Salesky was happy to try the new role. The software team was well staffed, and hardware—how sensors, cables, circuit boards, and the like all fit together in a rolling robot—presented plenty of problems to be sniffed out and squashed. The job took on extra importance when Urmson decided to build a new vehicle for the self-driving age.

The team had originally wanted to build its system into conventional cars that would drive themselves in some cases and let the human do the work in others. They soon realized that people and robots don’t play together nicely. After initial periods of nervousness, Googlers invited to test the tech got comfortable, and then too comfortable. Urmson watched them play on their phones, dig around the back seat, and fall asleep as they galloped down the highway at 65 mph in prototype machines. He and his teammates realized that the only safe way to deploy their tech was to take it to its logical conclusion: eliminate the human as driver.

A newly configured vehicle dubbed Firefly embodied that ambition. No steering wheel. No pedals. It was fully electric, with a 25 mph top speed and an exterior made of foam to mitigate the consequences of any crash. But its soft edges and polished exterior hid the stressful work that had gone into it. That foam exterior had proven onerous to shape and a pain to paint.

The limits of Google’s software meant operators still needed a way to take control, so the team had to hack in a flat metal disc that spun left or right to steer, not unlike Disneyland’s teacups ride. The vehicle didn’t have an HVAC system, so operators working chilly nights had to keep the windows open, lest the windshield fog over. The team never much used the Firefly platform and retired it in 2017. The lesson was clear: This sort of problem-solving wasn’t worth it. The work of building vehicles was best left to the automakers.

Starting in the fall of 2015, the Google team transformed. Page and his Google cofounder, Sergey Brin, hired former auto exec John Krafcik to head the effort as it attempted the move from developing the tech to making money off of it by launching a robotaxi business. Around the same time, core team members received six- or seven-figure bonuses, their reward for having stayed on the team for four years (the upshot of an unusual compensation scheme).

The leadership change, the cash that made a steady job unnecessary, and long-simmering tensions between teammates led to an exodus. Over the course of 2016, Urmson, Levandowski, and half a dozen other long-serving engineers left Google, nearly all of them to start their own self-driving companies. Their timing was good. The tech and automotive worlds were waking up to the lucrative and disruptive potential of the tech Google had pioneered and were pumping out cash like a blue whale’s heart pumps blood.

Salesky made his move in the fall of 2016. Because he left Google and returned, he wasn’t eligible for the four-year bonus payment but says he made enough money that he could have retired then. But he wanted to keep working, this time as his own boss, free to do things his own way. He was also tired of Silicon Valley. “I’m a four seasons kind of guy,” he says. By chance, his time in California almost perfectly matched the state’s 2012 to 2016 drought. He’d always kept a house in Pittsburgh, and now he moved back full time. The first time it rained, he walked outside and stood in the storm.

Among the companies chasing autonomous capability in this time, Ford should have had an advantage. It was the rare automaker that competed in the Darpa Challenges, fielding a team of its own engineers rather than sponsoring a university effort. The engineers posted respectable results in the 2005 and 2007 competitions but kept a low profile, naming their team Intelligent Vehicle Safety Technologies. The idea was to investigate the potential of self-driving technology, not attract attention to an experimental project. But where Larry Page came away from the Urban Challenge with a revolution in mind, Ford followed up with years of halfhearted research, aimed at making minor improvements to its human-driven machines.

Google, of course, didn’t have to worry as much as Ford did about the Great Recession, which, combined with lackluster products and locked-in union contracts, sent the automaker to the brink of bankruptcy. CEO Alan Mulaly restored the company to health with a focus on good governance and consumer-focused products. But his efforts to right a listing ship didn’t account for the tsunami on the horizon. The combined effects of ride-hailing, electric propulsion, and autonomous driving looked to remake the way people used their cars—and the workings of the folks who made them.


This was hardly the first time in its history that Ford was deaf to a changing marketplace. Henry Ford effectively created the auto industry with the Model T, which at one point gave him 60 percent of the US market. For the next two decades, though, he refused to design another car: The Model T was just fine, he believed, even as sales slumped and consumers flocked to a newcomer called General Motors, whose offerings got better year after year.

Ford’s only son, Edsel, convinced his father it was time to retool in the late 1920s. Ford finally produced the very popular Model A, but not before laying off thousands of workers amid the Great Depression and losing its dominance for good. A generation later, the company waited a decade to respond to Chevrolet’s Corvette with the Ford Mustang. This decade, it has lagged behind on electric cars, even as regulators around the world move to broaden the supply of vehicles beyond gas-powered cars and trucks.

When Mark Fields succeeded Mulally in 2014, he saw the change coming and determined that Ford would get ready for an age when being a good automaker might not be enough anymore. The shift away from the human driver threatened Ford’s core business. “The self-driving car wasn’t an enhancement of the existing technology,” John Casesa, who was then Ford’s head of strategy, said in 2018. “It was a substitute for it. It was like going from the horse to the car.”

At CES in Las Vegas in January 2015, Fields gave a keynote address declaring that Ford was no longer just an automaker. It was now, also, a “mobility company.” He opened a Silicon Valley Research and Innovation Center and hired scores of software engineers. More important, that spring, he started talking to Google’s self-driving team about collaborating by combining Ford’s hardware capabilities with Google’s software knowhow.

By December, the press was full of rumors about the deal. Later reporting by the Silicon Valley Business Journal and Automotive News portrayed Fields as eager for a high-profile deal that would impress investors and show he was serious about modernizing the aging automaker. But in January, Google walked away and instead made a deal with Fiat Chrysler to use its Pacifica minivans as its robo-mules.

When Fields returned to CES in 2016, he put on a brave face and announced that Ford would triple the number of cars it used to test its own autonomous driving tech, from 10 to 30. Eight months later, he appeared at Ford’s Silicon Valley outpost with a significantly bolder announcement: Come 2021, Ford would launch a self-driving ride-hail service in a US city, with at least hundreds of vehicles that wouldn’t have steering wheels or pedals.

The problem was that, in 2016—more than a decade after its engineers competed in Darpa’s Grand Challenge—Ford had no idea how to build a self-driving car. And if Google still hadn’t launched such a service after seven years of work, Ford stood no chance of figuring it out by 2021.

Casesa, who spearheaded this part of the business, knew Ford needed more than fresh talent. It needed a different way to structure the project and insulate it from the bureaucracy of a business that works on decades-long game plans. It needed the ability to attract top-quality engineers by offering equity and bonuses, and it needed to spend lavishly with little guidance on how much money it would make back, or when. “We weren’t a software company,” Casesa says. “So we were going to create a software company.”

Not long after his return to Pittsburgh, in the fall of 2016, Salesky sent a note to his old NREC boss and friend, Peter Rander, saying he had some news to share—he had left Google. Rander replied with his own, similar news, that he too was becoming a free agent. After 14 years with NREC, he had in 2015 helped start Uber’s self-driving program. CEO Travis Kalanick had been eager to catch up to Google on tech that promised to make Uber profitable—or, in a competitor’s hands, bankrupt it—and so he lured away dozens of NREC engineers, including Rander, who joined because the company had the chance to build a product that could improve the lives of millions. (The money was a whole lot better, too.)

The team struggled to meet Kalanick’s aggressive goals and after a year was subjected to a structural shake-up: Kalanick had acquired the self-driving truck startup Otto and put its leader, Anthony Levandowski, in charge of Uber’s autonomy effort, then called the Advanced Technologies Center. Like Kalanick, Levandowski liked to move fast, and he wasn’t impressed with his new employees’ work. “Wow,” he texted Kalanick. “I am super pissed at what is going on at ATC. There’s one one [sic] who is pushing for the right things.” New management didn’t improve things much, and the effort took a morale hit when Waymo sued Uber, accusing it of hiring Levandowski to access a trade-secret-filled trove of documents he had downloaded on his way out the door.

So after a bit more than a year, Rander cleared out his desk. He met Salesky for dinner outside Pittsburgh at Luciano’s Italian Brick Oven, a cavernous restaurant done up as a Roman villa. Working off a menu heavy on pizza, calzones, and hoagies, the two caught up and started batting around ideas for things they could do together.

Soon, they settled on self-driving. They knew the tech and the business landscape, and they saw an untapped opportunity. Salesky had been in on Google’s negotiations with Ford and had seen how the auto industry was hungry for autonomy but ill-equipped to deliver it. (Ford’s crosstown rival, General Motors, had made its move in early 2016, acquiring the self-driving startup Cruise to replace its in-house effort.) And he knew that when it came to building complex, durable vehicles at scale, no tech venture could compete with Detroit. Argo would be the yin to the auto industry’s yang: sharp on software but letting the manufacturer handle the manufacturing of vehicles. “I didn’t think that entity existed,” Salesky says. “So we put that hook in the water.”

Waiting in the deep was John Casesa, who by late 2016 knew Ford needed outside help if it was going to come anywhere near meeting Fields’ 2021 deadline. He had met Salesky during Ford’s negotiations with Google the year before and liked him. Together, they created an unorthodox setup, especially for a Detroit company used to outright acquisitions. Argo would be an independent startup, with its own board and the right to offer equity to employees—a key recruiting tool in a field where engineers are scarce. It would be able to sell its software to others, with board approval. (For its $1 billion investment, Ford would get a major stake in the startup and appoint two of five board members.) And while Ford’s public relations team insisted the effort would complement research done in Detroit, the move effectively ended Ford’s own autonomous-driving research effort. Argo invited some Ford workers to interview for jobs.

As they built their core team—the folks to lead a staff that would grow to more than 200 by the end of 2017, and more than 800 today—Salesky and Rander sought out engineers who, like Rander, had left NREC for Uber and were ready to leave the troubled company. They included robotics chief Brett Browning, autonomy integration lead Al Costa, software engineer Randy Warner, and half a dozen others. Twelve years after Salesky had walked into NREC, met its robots and roboticists, and decided that’s how he wanted to spend his life, he had, in a way, rebuilt the team. But this time around, he was the man in charge of keeping them on the right heading.

Three years on, Argo—Salesky thought the name was cool, and the web domain was available—has made its headquarters in a square building of brick and glass in Pittsburgh’s newly hip Strip District, less than a mile and a half from Uber’s self-driving research center and the NREC office. More than 300 software engineers fill two floors, many of them at desks that electrically rise to standing height or huddled in conference rooms named for local neighborhoods. They enjoy free catered breakfast and lunch, plus snack containers retrofitted with handles engineered to ease access to M&Ms and Chex Mix. Take away the view of the Allegheny River on one side and the abandoned railroad tracks on the other, and this could be any Silicon Valley office. (The company does have an outpost in Palo Alto.)

When I visit on a Tuesday in October, it’s raining, and the resulting traffic makes Salesky a few minutes late—he drives in every day from about 40 minutes away, usually in a Ford F-150 pickup. I’m eating hash browns and playing with my phone when I look up to see him standing over me, in sneakers and an untucked button-down, with a big smile. He’s quick to laugh and happy to tease his PR handlers.

He tells me that he’s confident he’s made the right kind of deal with Ford, giving his team access to a manufacturer without the responsibility of making a vehicle themselves. When he needs to adjust the car’s regenerative braking, he just makes a phone call. “It’s super powerful to be able to say, ‘Hey we need to change the shape of the vehicle in order to optimize sensor field of view,’” he says. “Or ‘Boy, I’d really love to be able to talk to the folks who do headlamps.’” Especially since those are the people who know what one small change can do to the rest of the system.

In his NREC days, Salesky and his colleagues would casually pull the spare tire out of the trunk and pile their electronics in its place. Once, an automotive safety expert came by and told him that was a bad idea—in that particular model, the spare tire was a key component in the car’s crash performance. “A bunch of software guys didn’t think about that,” Salesky says. Now, they don’t have to.

What they do have to think about is how to produce the code that will allow a car to drive itself. Progress there is hard to measure. The system performed admirably in rides I took in Pittsburgh and Miami, where Argo and Ford hope to launch a ride-hail and goods-delivery service, but the real test is how they work over millions of miles. That will come down to the work of the hundreds of engineers Argo has recruited over the past two years, and thus to Salesky’s leadership.

Here, the CEO has his detractors. Some former Google colleagues charge him with a lack of charisma, which they deem a handicap when it comes to motivating a team to overcome a brutally difficult obstacle, even if they don’t have to worry too much about 2021. Mark Fields’ pledge doesn’t come up much these days: He was ousted from Ford’s top office in May 2017, and Salesky never committed Argo to the deadline. In late April, Ford pushed the delivery date to 2022, citing “the challenges of the current business environment, as well as the need to evaluate the long-term impact of Covid-19 on customer behaviors.”

It’s true that Salesky lacks the magnetism of some tech executives. In his rare public appearances, he speaks quietly and tends to go deep into technical details. Unlike some competitors, Argo doesn’t tout how many miles its vehicles have driven and only rarely invites the press to take a ride in its cars.

But Salesky bristles at the suggestion that being well-known equates to being well-suited to this challenge. “I am an average Joe from Detroit. That is who I am,” he says. “I don’t seek out press. I don’t seek out headlines. I don’t need that validation. This is why you’ve never heard of me.”

His reluctance to talk about his personal life or to fashion an inspiring origin myth for himself clashes with the Silicon Valley ethos, where many a CEO’s job description includes boosterism. Think Elon Musk, Adam Neumann, Travis Kalanick, Elizabeth Holmes. They shine (or shined) in the spotlight, winning over investors and turning consumers into fans. But the struggles of Tesla, WeWork, Uber, and Theranos make clear the difference between pitch and delivery.

“I’ve been involved in the industry for a long time. I’ve had a great impact,” Salesky says. “Not just academic research that gets talked about in papers, but actual products that get delivered.” As a result, he doesn’t have to worry about selling his vision. He has Ford behind him, and as of July had won over another major partner: Volkswagen. Linked to a broader alliance between the automakers, VW is investing $2.6 billion in Argo and merging its own Autonomous Intelligent Driving subsidiary into the startup. The companies say the deal values Argo at more than $7 billion, and perhaps more importantly it validates Ford’s early bet on the company. VW had previously been working with Aurora, the self-driving company run by Salesky’s old friend, Chris Urmson. And its in-house effort was more robust than anything Ford had mounted. But now the German giant is leaning on Argo to produce the technology that might drive its future.

Curiously, the best place to get a glimpse of that future isn’t Pittsburgh or even Silicon Valley. It’s Miami, where Ford and Argo are hoping to launch a robotic-taxi and delivery service in the next few years. The partners took over a 20,000-square-foot warehouse in the Wynwood neighborhood. To match the area’s famed murals, Ford hired a local artist to paint the building in a mildly psychedelic blast of orange, blue, purple, and green, where cyclists mix with a flamingo in flight, orbs sluicing along road-like ribbons, and a telescoping eyeball. The theme, I’m told, is “freedom for all.”

The warehouse’s parking lot, though, is the more interesting bit. Behind a tall, opaque fence sits Argo’s fleet of test cars, each gleaming white and blue and meticulously clean, even if no one from the public is welcome inside yet. Here I meet Rander, who joins me for a ride. Safety operators Christian and Felipe sit up front and chatter constantly, calling out everything they see—a technique that keeps them focused on their surroundings.

The car rolls by scooters no problem; it leaves plenty of room for the potentially confusing sight of a cyclist holding a second bicycle as he pedals past. Then it brakes harder, apparently to build up an unnecessary amount of space between itself and a slower car in front. Felipe starts typing on his laptop. He’s noting the behavior, Rander says, not because the car broke any laws or endangered someone but because it breached the social contract of the road. Lots of cars around here brake occasionally, as their drivers take in the murals. The trick is to give them enough space without clogging traffic in your wake. “We need a ride that’s comfortable, confident, and intelligent,” Rander says. “I don’t mean rocket science. I need common sense.”

That kind of metric—more than miles, funding dollars, or public attention—is what Salesky keeps his eye on as Argo marches toward launching its technology, in 2022 or whenever it’s ready. His role as CEO, ultimately, isn’t so different from what he did for Carnegie Mellon’s Urban Challenge team. He spends his days talking to his managers, checking their progress against the goals they’ve set. He spends lots of time with the operations team, talking to the test drivers who live with the cars every day. The big difference now is that he, the quiet American, is the one in the driver’s seat.

Related Articles

Latest Articles