Full confession: I've taken a drone selfie.
I did it the other day, in the afternoon sunshine along a scenic stretch of San Francisco Bay. I was holding a smartphone, my arm extended like a wizard casting a spell, and I kind of felt like I was. I raised my hand up and to the right. The small quadcopter drone buzzing a few yards in front of me followed. Then I waved my arm in a sweeping arc to the left. Again, the drone obeyed as cameras mounted on the aircraft recorded my moves.
To land, all I had to do was swipe down on the screen, the same gesture I use every day to view my notifications.
Despite my success, I am in no way a pilot. But Adam Bry is. As a kid, Bry was a radio-controlled airplane fanatic, a passion he channeled into multiple victories in national aerial acrobatics competitions. He went on to study computer science as a grad student at MIT before landing at Google, where he was a co-founder of the company's formerly secret drone delivery research program, Project Wing. Co-founding a drone company was the obvious next step.
>The hunch is that the real money won’t be in drones themselves, but the software that makes them useful.
Today, that company, Skydio, announced its first infusion of serious money, a $3 million seed round led by Andreessen-Horowitz. (Bry is the CEO; fellow MITer and Project Wing co-founder Abe Bachrach is the CTO.) The money will go toward bulking up Skydio's effort to build true computer vision into drones, enabling them to navigate based not on GPS but on what they "see."
With a few rare exceptions, drones in the US still remain grounded by federal regulators---at least those drones intended to be used for commercial purposes. But that hasn't stopped Silicon Valley investors from flocking to so-called unmanned aerial vehicles (the nascent industry's preferred term), pouring about $95 million into drone startups over the past two years. Investors are betting widespread use of drones is only a matter of time as government and cultural norms catch up with the technology. And when they do, the hunch is that the real money won't be in the flying machines themselves, but the software that makes them useful.
Some well-funded startups, such as Skycatch and Airware, see drones as one of the next big computing platforms. They're developing applications for businesses to take advantage of drones' unique ability to see the world from above for a small fraction of the cost of manned aircraft. If heavy industries such as construction, mining, and agriculture can have eyes in the sky all the time, the thinking goes, they can gather more kinds of data, and more of it, than ever before. In that context, the value is not in the drones themselves, but rather the insights they provide.
Skydio is trying to solve a problem further down the stack, to put it in tech industry jargon. Rather than working on specific applications for drones, the company is trying to make drones much better at getting around on their own. Right now, Bry says, drones largely depend on GPS to identify their locations in space. GPS lets drones fly to specific locations without human intervention, but isn't much help when it comes to avoiding unexpected obstacles they might encounter along the way.
>Skydio's co-founders believe you shouldn't need a pilot's understanding of the mechanics of flight to find uses for a drone.
The answer, Bry says, is cameras powered by software that recognize obstacles as they appear. It's that software, fused with the right lightweight hardware, that Skydio is working to build.
"You can think of it like a self-flying vehicle instead of a self-driving car," Bry says.
The ultimate intention is to make drones as easy to fly as opening an app. If flying a drone becomes trivially easy---in other words, if it can be done by someone like me---then drones become accessible platforms for a broad range of endeavors. Instead of just flying over a construction site, for example, the drone can fly inside the unfinished building.
Or an aerial ignoramus like me can use a drone to take a picture of himself. You don't have to know how a smartphone works to find a multitude of uses for it; in the same way, Skydio's co-founders believe you shouldn't need a pilot's understanding of the mechanics of flight to find uses for a drone.
"We think of control sticks like the command prompt," says Skydio co-founder Matt Donahoe, who met Bry and Bachrach while at MIT's Media Lab.
Skydio is not the only company working to give drones the power of sight. In one of the most popular demos at the Consumer Electronics Show lasts week, Intel CEO Brian Krzanich played "drone pong", a game in which an obstacle-sensing drone hovering over the stage moved away as he approached. Regardless of who gets there first, the brainpower being applied to the problem suggests a solution is only a matter of time.
When that future arrives, drone enthusiasts promise a flourishing of creativity as another technology that, like the computer, was once limited to experts becomes accessible to everyone. But many members of that broader public that would make up the market for democratized drones are still deeply suspicious of potentially prying airborne eyes. For the pro-drone contingent, overcoming that mistrust could be the most important thing easier-to-use UAVs can accomplish. After all, what could be cuddlier than an airborne selfie stick?
