Fernando Cwilich Gil exploited a bug in Apple’s OS to create "Martine, the girl who doesn’t know whether she’s coming or going."
Anthony Ferraro’s "Hypothetical Beats" sold at the Algorithm Auction for more than $10,000 in 2015.
In 2013, a painter named Fernando Cwilich Gil stumbled upon an idea while living in his own world—a private island in Brazil. He had been commissioned to produce one of his surrealistic landscapes to hang in a patron’s house in Angra dos Reis Bay, Brazil’s version of the Hamptons. And as he set to work on the piece, taking preliminary photographs with his iPhone, he accidentally made a discovery: Programming code doesn’t just run our tech—it can be the basis for art.
“It’s a long, strange journey,” Gil laughs as he works on a new canvas in his Brooklyn studio on a blustery winter day, far from the Brazilian island where, when not painting, he fished for dorado. The artist recounts the moment of his epiphany, when he discovered that by moving his iPhone in a certain light with a certain undulating motion (“a combination of sputtering and rocking of a janky Brazilian fishing boat,” he says, “with the drawing of a net, the casting and tugging of a fishing line”), he could take pictures in which people appeared with reversed heads or erased bodies. “I discovered you could alter the pictures without using Photoshop,” he says. “I thought everyone knew about it.”
They didn’t. Gil showed the bug to a friend on an ensuing trip to the United States, and “it blew me away,” says Ben Gleitzman, a programmer based in Oakland. Gleitzman and some colleagues examined the code that ran the camera on Gil’s iPhone 4S and found that he was exploiting what the engineer calls “a deep bug in Apple’s OS”—three separate algorithms that didn’t quite work properly together. Some of Gil and Gleitzman’s friends tried to talk them into forming a company to sell a photo-manipulation app that used this discovery, but Gil balked. This wasn’t a commercial breakthrough, he argued; it was an artistic one. The algorithms themselves represented “museum-quality code,” he says. It was the digital building blocks he found beautiful, not just the images they created.
So, with art world puckishness and Silicon Valley pluck, they formed a small company called Ruse Laboratories that, in addition to throwing elaborate parties (Gleitzman had to duck out of our interview early to pick up a Japanese DJ from the airport), would allow them to convince the rest of the world that code could indeed be treated as art. After New York’s Museum of Modern Art gave them a cold reception (a curator told them the Facebook interface would never make the MoMA because of the “ugly” blue thumb), the Cooper Hewitt, Smithsonian Design Museum, which had embarked on a digital-heavy renovation in 2014, welcomed them. On March 27, 2015, they held the Algorithm Auction, a splashy sale of representations of programs like Hello, World! and the OkCupid match algorithm. They sold the pieces for thousands of dollars each (proceeds went to the museum)—and garnered coverage in the Wall Street Journal, the New Yorker, and Wired.
Today the pair are connecting the programmers who will build future “museum-quality code” with funders willing to patronize them. They call the program Mecenas, after the Spanish word for patronage. With Ruse acting as a mediator, the way an art gallery might, in the last year wealthy patrons have funded more than 20 projects, some at a few hundred thousand dollars at a time. Coders retain the intellectual property rights, and funders buy an option for commercial development of the artwork, albeit with the understanding that the coder’s original intent must be maintained within the final product. (Ruse wants to support integrity of ideas, says Gleitzman, not yet another “Tinder for dogs.” But this is the tech world: Everything is negotiable.)
The idea that code could be art didn’t originate with Gleitzman and Gil. For decades, programmers have borrowed from mathematics to talk about beautiful code. Beginning in the 1980s, MIT computer science professor Hal Abelson began his introductory lectures by writing “Computer Science” on a blackboard and crossing out both words. Code isn’t always empirical, he would say—it’s closer to magic. In 2003, venture capitalist Paul Graham published an essay online titled “Hackers and Painters,” in which he argued, “Of all the different types of people I’ve known,” those two vocations “are among the most alike.” Programs, wrote Greg Wilson in the 2007 textbook Beautiful Code, “could be as elegant as well-made kitchen cabinets, as graceful as a suspension bridge, or as eloquent as one of George Orwell’s essays.”
But wait. How can you auction a piece of math? “It’s ephemeral. It’s not something you can touch, like a Greek vase,” says Gleitzman. In the case of the Algorithm Auction, they sold memorabilia—a framed pencil-and-ink doodle of OkCupid’s match algorithm; a signed page with two lines of C that printed “Hello, World!”; a tie featuring the code that unlocks DVDs keyed to specific regions. It’s no different, Gleitzman says, than auctioning off a print of a photo without the negative, or a recording of a dance performance. He laughs: “They’ll sell Duchamp’s toilet, but they won’t sell these thousand lines of code?”
But how to display code to an audience that’s code illiterate? You don’t have to understand the C-major scale to listen to Mozart’s Ninth Symphony, but don’t you have to know C++ to appreciate code? For now, Gleitzman and Gil focus less on technical explanations and more on expository ones. What problem was the programmer trying to solve? How did they do it? In this way, curation becomes explanation. That may change, though, as coding literacy becomes more widespread. “I was in L.A. at a friend’s house at a kid’s birthday,” Gleitzman says. “All these little kids were arguing about computer languages. They were talking about it in the way somebody would talk about French or Spanish—or paint or clay. We’ve been taught that code was something utilitarian, and not thrilling. That’s a problem.”
OK, so let’s stipulate that code can thrill. But does that make it art? Just because art is auctioned doesn’t mean that all things that are auctioned are art (see: Beanie Babies). Further, isn’t it awfully problematic to say, in a time in which coders are displacing artists across the Bay Area, that code ought to displace traditional forms of culture, too? Steve Lambert, who teaches in the new media program at SUNY Purchase, rolls his eyes. “That’s how bad it’s gotten [in the Bay Area],” he says. “All the artists have been driven out, so maybe all these people in coding, they’re artists. That’s how desperate we are for culture.”
But, say Gleitzman and Gil, there’s less distance between coders and painters than you’d think. “If you were a Flemish painter in the 16th century, you were a nerd,” argues Gleitzman. (And, true enough, it’s not so hard to imagine Bruegel kitted out in a hoodie and hunched over the command line on his laptop.) To determine whether code can be art, we’d have to know what art is, anyway. And good luck with that. “Since World War I, it’s been understood to be a waste of time to debate whether [a medium] is art or not,” says Renny Pritikin, chief curator at the Contemporary Jewish Museum. “The question is whether it’s successful.”
And here’s where, maybe unexpectedly, Gleitzman and Gil find themselves on solid ground. For at least 50 years, artists have been working with written instructions as a medium. In 1961, Yoko Ono created an “instruction painting” inviting viewers to hammer a nail into the art. In 1968, Sol LeWitt embarked on a series of wall drawings in which he gave instructions like “On a wall surface… place fifty points at random” for others to execute. So, ask Gil and Gleitzman, what’s the difference between an early Ono and an early Zuckerberg?
It depends. Art has many different meanings, says Vikram Chandra, an amateur coder and writer who teaches at UC Berkeley. “What do we mean? Art as opposed to science? Art as something that is practiced by a practitioner of great skill? Like a carpenter’s art? Or fine art?” Some code can be art, he thinks, but not all of it. It would have to be written from a place of free imaginative play, rather than to run a grocery store website, and it would have to solve a problem so elegantly as to provoke a sense of wonder. “To casually claim that code is art is a desperate misunderstanding of the life of an artist. That irritates me immensely,” he says.
Gil knows that what he’s doing has rubbed some people the wrong way—an essay he wrote for Artsy a few months ago provoked a big pushback. That doesn’t shake him: “We’re in the vanguard, and we’ve pissed off a lot of people in the art world.” Once upon a time, the murals inside Coit Tower provoked as strong an antipathy, he says. “If I had to set out to make today a piece of art that would have a significant social impact,” he says, “I’d make code. Not a mural.”
Originally published in the May issue of San Francisco