Self-driving cars are already rolling along in Pittsburgh, thanks to Uber (albeit on a small scale with humans onboard, ready to intervene), and a Wired writer gave it a shot. A bevy of companies are working to put autonomous cars on the streets, but a new announcement by Udacity at TechCrunch Disrupt SF could and should send shockwaves into the nascent industry.
Udacity is best known as a titan of online education, specializing in "nanodegrees" for people interested in working in the tech sector. For $2400 and a 9-month commitment, Udacity can turn prospective students into viable experts on self-driving vehicle technology, capable enough to work with the likes of Google, Uber, and other firms working on this next step forward. Of course, new students will need a background in programming, but the course will offer the chance to master deep learning, sensor fusion, vehicle kinematics, and more subjects to enable your new Tesla drive on its own accord.
"We will be the only institution, I believe in the world, that would give a credential in this space," Udacity CEO & Cofounder Sebastian Thrun said. "We decided at Udacity we wanted bleeding edge."
The most significant part of Thrun's announcement was that open source the whole project; everything that comes out of their processes will be freely and publicly available for anyone to access, copy, and fork for their own uses. The e-learning colossus has plans to put autonomous wheels on the road by year's end and announced a slate of top-tier partners for the new program: Otto, a self-driving truck startup acquired by Uber in August, will lead on hardware; Mercedes-Benz will deploy their North American R&D arm to guide software; Nvidia will give the curriculum an overall direction; and Chinese ride-sharing giant Didi will host machine learning competitions for Udacity in exchange for a sneak peek at the nanodegree students' projects.Udacity will
While the juggernauts of Silicon Valley race to make their own autonomous vehicle offerings, Udacity's move to open source whatever comes out of their new program has incredible potential to disrupt a disruption. With all of that tested and reviewed code freely available for anyone to use, the barrier to entry for the self-driving car market will suddenly be lowered and the game will be open to more players. Given that this is Udacity's first dive and it's with students, their initial product may be incomplete, but it still gives the masses a starting point for building.
That is, unless one of the big players takes a platform approach to the self-driving market.
We know the future of the automotive industry; the supply chain will never be the same again. In order to win and conquer this evolving streetscape, the ripest opportunity isn't in the factories making the engines or even the offices churning out the code needed to build the AI. Domineering success will be had by anyone who can build the first viable platform for connecting app developers with consumers.
With their hands and attention now freed from driving, motorists are now passengers in their own vehicles, free to watch Netflix and Hulu, check in on social media, and even knock out some work on their commute to the office. This new source of activity will require an interface and a marketplace chock-full of apps to occupy their time. In other words, an operating system coupled with an adapted version of Google Play or the App Store built right into their vehicles.
The advent of the automotive app has incredible potential, beyond simply mimicking the capacity and function of smartphone apps with a larger screen. Users could make dinner plans while on their way to their destination, order food from a drive-thru much sooner and avoid the wait. Apps could even show passengers the live state of the car and provide a dynamic repair schedule.
A wise move by Udacity would be to create this development platform themselves (or encourage their students in that direction) along the way and open source it, much like was done with Android. Offering licenses of this new platform to manufacturers of autonomous vehicles (think the Big 3, who have historically lagged in innovation) would enable Udacity to leapfrog to monopoly status in a market of their own creation and define the future for automotive apps.
With nearly 8,000 application vying for 250 seats as of this writing, Udacity has clearly hit on something great, but it remains to be seen if they or anyone else understands the platform potential they've introduced by open sourcing their self-driving car program. If a self-driving company doesn't move quickly, they could find themselves outdone by a network of open source developers or another firm that exploits such a network.