Google Earth may not be as useful as its sibling service, Google Maps, but it’s immeasurably cooler. Maps take you where you need to go; Google Earth takes you where you want to go—not just to some dumb restaurant but to far off lands and unreachable peaks and even to the ocean’s floor. It’s all so detailed and fluid that you can forget what it is, exactly, you’re looking at—not our planet but an incredibly complicated 3-D model of it, assembled entirely by algorithm and code. As we see here, sometimes that reality shows through.
These upside-down bridges and drooping motorways are from "Postcards From Google Earth," a collection of screenshots by Brooklyn-based artist Clement Valla. Even on their own, plucked out of the program, they’re a little unsettling. We expect Google’s Earth to look like the one we see out the plane window, with straight roads and solid overpasses. Here all that infrastructure looks like undercooked pappardelle.
You might be tempted to call them glitches. But that’s not quite right, and in part why Valla finds them so fascinating. It helps to understand what’s going on. The effect occurs when the algorithms that piece together Google Earth encounter aerial photography with too much depth, or too extreme an angle, or too many shadows. So these gelatinous bridges aren’t the result of the system breaking down, Valla explains, they’re things that are "inherent in the system itself."
"It just happens to be sort of an edge condition," he says. "That’s what I like about these. The computer’s doing exactly what it should be doing."
For Valla, whose work often deals with automated systems and finding ways to draw attention to them, the warped patches were an exciting find. "I like to expose algorithms," he says. "Most of the time that can be a cold, intellectual pursuit. I think these are successful because there’s a visceral reaction. You almost can’t help but be like, 'whoa, that doesn’t make sense.'"
From there you can get into the bigger questions they raise about algorithms and computers and automation. Which might be worth thinking about. Some day soon, Google’s systems could be responsible for driving our cars. And you have to wonder what their edge conditions might look like.