By Chris Urmson
Few things prompt you to revisit the path of your career more than a documentary that tracks the narrative you’ve been living since graduate school. Autonomy, directed by Alex Horwitz and produced by journalist Malcolm Gladwell, premieres at SXSW this weekend. In the film, I, and others, recount our experiences helping to develop self-driving technology. The film asks great questions about what our future with self-driving cars will be like, and when we’ll get there. (It’s great, you should see it.) Coming on the heels of Aurora’s recent financing round, and the meaningful progress we’ve been making in our development, I thought it made sense to step back, and assess this moment — as well as the industry in which I’ve been fortunate to spend my career.
Recently, it’s become a bit of a meme that self-driving vehicles are over hyped, or worse yet in the “trough of disillusionment.” Maybe. And maybe something like this was inevitable. Some of our industry’s leaders could have done a better job of setting expectations about the difficulty of the problem. On the other hand, none of us truly knew how hard delivering this technology would be.
In reality, it’s difficult to say we’re in a hangover phase when our industry continues to attract the sort of funding ours does. Many companies in the sector have been the beneficiaries of significant investments. Simultaneously, we’re also starting to see consolidation happen, which is normal in any nascent industry. Building self-driving vehicle technology is challenging, absolutely, but the scale of these investments reflects the widespread confidence that it’s solvable — and that the impact will be enormous.
Watching Autonomy, I marveled at this community’s journey. It was fantastic to see the early pioneers, such as Sadayuki Tsugawa and Ernst Dickmanns, being recognized for their groundbreaking work. It was equally great to briefly revisit the DARPA Grand Challenges, the training ground for many of the leaders of the self-driving industry today. A good decade before the challenges, my predecessors at Carnegie Mellon traversed the continental United States in 1995 in a Pontiac Trans Sport minivan equipped with a lane-keeping computer program. They called the trip “No Hands Across America,” because, while the software steered the vehicle, the team had to apply throttle and brake.
That achievement was incredible at the time.
Since then we’ve made enormous amounts of progress. The industry has matured from a handful of academics working on a few college campuses to a global industry populated by fast-expanding companies. More personally, in just the two years since founding Aurora with Sterling and Drew, we’ve grown to more than 200 people and operate our vehicles daily on the roads of San Francisco and Palo Alto, in California, and Pittsburgh in Pennsylvania.
Sterling, Drew and I founded this company to deliver self-driving technology safely, quickly and broadly. The safely component is intrinsic to every step of our development. Our technology will ultimately be safer than a human driver. Quickly is predicated on the safely. Getting this technology right is urgent since car crashes around the world take the lives of two and a half people every minute.
Now for broadly, the key part of our business strategy. When we started our work together, Sterling, Drew and I did what most engineers do when they start a new project. We asked, “Where can we have the most impact?” Each of us had gone through similar processes at projects within companies, or on academic faculties. Now we were assessing, how can this company we’ve created contribute most to society and our future?
The autonomous car sector did not lack for ambitious business models. Some hoped to invent autonomy even as they scaled ride-sharing businesses across the world. Others were hoping to reinvent vehicles, plus innovate ride-sharing, and on top of all that, invent autonomy. We saw an opportunity to solve this more efficiently if we focused our ambition on the component we were uniquely qualified to deliver — the self-driving technology.
What Sterling, Drew and I felt our team, together, could do better than anyone else was to develop the Aurora Driver — the hardware, software and data services that vehicles require to pilot themselves about the world.
We’ve done that in a particular way. Broadly — one of the things that means is that we’ve designed our technology from the beginning to be a platform. We intend it to operate, as the driver, across vehicle makes and models, from the largest tractor-trailer or passenger bus to small cars and anything else that drives on the road. It can operate in the service of moving people and goods, as part of transportation networks, logistics services, or for last-mile delivery.
As a platform, our independence is important for our success, and the success of our partners. It also means we’re not beholden to exclusive partners or their products. Independence means we have our pick to create relationships with the best in the business — for every part of the ecosystem. Which in turn makes it easier for us to develop the safest technology, and increases the likelihood that we’ll be fast to scale.
When will the technology be ready? That’s a difficult question, but we expect the first small-scale deployments of self-driving vehicles will occur in America within the next five years. We’re moving quickly.
I think back to 2003, as I was sitting on a folding chair on a salt flat in the Atacama, making tweaks to software code in the hopes that I could help get our rover to drive at 15 cm/s. Had you shown me then the things we’re able to do today, I would have been amazed. Maybe it’s taking a little longer than we thought — but the future of transportation is looking a little brighter, and a little better every day.