It's easy to poke fun at Google's design for its autonomous car system, especially after the reveal of its dorkly-looking prototypes last week. But the scope of Google's autonomous car venture is larger than most people may realize. When we think of self-driving cars, we first thing about the car itself, and the technology required to give the car real-time awareness of its surroundings. But those LIDAR scanners mounted to the top of the cars are only half of tech. The other half is the wholly new infrastructure required for these cars to tap into Google's network and data. These cars will likely use cellular to access traffic and mapping data, but what networks will that tap into? Presumably not just into existing commercial wireless networks that smartphones use, or at least not part of users' private data plans.
Additionally, as this Atlantic piece on Google's autonomous car technology points out, the cars are only able to navigate through areas in which Google has thoroughly mapped into a new database. This isn't just Google Maps or Street View--it's an "ultra-precise digitization of the physical world, all the way down to tiny details like the position and height of every single curb." And that may be the biggest challenge to getting autonomous cars on the road. For its testing, Google's cars and sensors have "crawled" about 2,000 miles of road (much of it in Mountain View, where the cars are currently being tested). But there are millions of miles of roads left to digitize--a challenge Google seems to be willing to take on. And that's where things could get scary. It's easy to see why Google would want to take tackle the herculean task of surveying every aspect of our roads into their databases--it's incredibly valuable data that could go far beyond the implementation of autonomous cars. What Google chooses to do with that data is completely up in the air; if data is the currency of the Petabyte age, we're giving Google carte blanche to literally mine the streets for that digital gold.