Open Source Code Will Survive the Apocalypse in an Arctic Cave
(Bloomberg Businessweek) -- The last stop for civilization before the North Pole is Svalbard, an archipelago north of mainland Norway along the 80th parallel. Most of Svalbard’s old Norwegian and Russian coal mines have shut down, so locals have rebranded their vast acres of permafrost as an attraction to scientists, doomsday preppers, and scientist doomsday preppers. Around Svalbard, things can be hidden from the stresses of the outside world. There’s a treaty in place to keep it neutral in times of war. In other words, it’s an ideal spot for a big global reset button or two.
Pride of place belongs to the Svalbard Global Seed Vault, where seeds for a wide range of plants, including the crops most valuable to humans, are preserved in case of some famine-inducing pandemic or nuclear apocalypse. The seed vault looks like something out of a movie, its entrance a triangular obelisk jutting high out of a blinding white expanse. It sparkles with glowing green lights.
Nat Friedman, however, hasn’t come for the beat-the-apocalypse aesthetics. On Oct. 24, the tall, thin, 42-year-old chief executive officer of GitHub Inc., Microsoft’s world-leading code bank, hops in a van and drives about 15 minutes from his hotel to an abandoned coal mine, where he puts on a miner’s helmet and headlamp. Deep inside one of the mine’s frigid, eerily quiet arteries, Friedman comes to what looks like a metal tool shed. “It’s more mine-y and rustic and raw-hole-in-the-rock than I thought it would be,” he says.
This is the Arctic World Archive, the seed vault’s much less sexy cousin. Friedman unlocks the container door with a simple door key and, inside, deposits much of the world’s open source software code. Servers and flash drives aren’t durable enough for this purpose, so the data is encoded on what look like old-school movie reels, each weighing a few pounds and stored in a white plastic container about the size of a pizza box. It’s basically microfilm. With the help of a magnifying glass, you—or, say, a band of End Times survivors—can see the data, be it pictures, text, or lines of code. A Norwegian company called Piql AS makes the specialized rolls of super-durable film, coated with iron oxide powder for added Armageddon-resistance. Piql says the material should hold up for 750 years in normal conditions, and perhaps 2,000 years in a cold, dry, low-oxygen cave.
Friedman places his reel on one of the archive’s shelves, alongside a couple dozen that include Vatican archives, Brazilian land registry records, loads of Italian movies, and the recipe for a certain burger chain’s special sauce. GitHub, which Microsoft bought last year for $7.5 billion, plans to become by far the biggest tenant. Eventually, Friedman says, GitHub will leave 200 platters, each carrying 120 gigabytes of open source software code, in the vault. The first reel included the Linux and Android operating systems, plus 6,000 other important open source applications.
Yes, this may seem like a stunt, headlamps and all. If the world is ravaged to the point where Svalbard is the last repository of usable wheat and corn seeds, the source code for YouTube will probably rank pretty low on humankind’s hierarchy of needs. Yet to Friedman, it’s a natural next step. Open source software, in his view, is one of the great achievements of our species, up there with the masterpieces of literature and fine art. It has become the foundation of the modern world—not just the internet and smartphones, but satellites, medical devices, scientific tools, robots.
The basic idea of open source is that you write code and share it, giving anyone else a chance to see what you’ve done, and, if they like, to take the code and change it and make their own thing. Over time this vast and expanding body of work is repurposed and improved upon and used to make innumerable software applications. GitHub is where much of the world’s open source software gets developed. About 40 million people, many of them volunteers, refine the projects, log bugs that need fixing, scan for security holes, and track changes. Think of it as a gigantic, meticulously cataloged library of tools that anyone can use.
Open source is the dominant procedure for software development, though it took a revolution to get there. In the 1990s, at the height of the Microsoft Windows empire, Bill Gates’s subordinates described the code-sharing model as “a cancer,” a threat to everything that patent-loving capitalists should hold dear. “If you told someone 20 years ago that in 2020, all of human civilization will depend on and run on open source code written for free by volunteers in countries all around the world who don’t know each other, and it’ll just be downloaded and put into almost every product, I think people would say, ‘That’s crazy, that’s never going to happen. Software is written by big, professional companies,’ ” Friedman says in the vault, which he describes more as a time capsule than a critical insurance policy. “It’s sort of a magical moment. Having a historical record of this will, I think, be valuable to future generations.”
To many in the software trade, the craziest and most magical thing here is a Microsoft executive extolling the importance of open source. The rise of open source has indeed been huge, epochal even. And, like many significant inventions—nuclear power, antibiotics—open source carries risks. Some pretty weird ones, it turns out.
Programmers freely swapped code long before Linus Torvalds wrote the core of the Linux OS at the University of Helsinki in the early 1990s, but his creation was a standard-bearer for what became known as “the free and open source software movement.” Microsoft was making obscene amounts of money through Windows and Office, and closely guarded the source code of these products. As the U.S. Department of Justice began trying to reckon with Microsoft’s influence over innovation and competition, DIY hacker types such as Torvalds argued that the very idea of patented proprietary software stood in opposition to free speech, free access to public goods and knowledge, and progress itself. (This was less radical than it might sound; U.S. law didn’t recognize software as intellectual property until the late ’70s.)
These idealists injected a dose of counterculture spirit into the debate over how much control a few large companies ought to have around technological advances. Linux became the most prominent alternative to Windows, and other coders created a free package of open source Office alternatives called, of course, OpenOffice. Both products struggled to find a mainstream audience, partly because the developers were sometimes more focused on the source code’s purity than on its usability. Yet they gained valuable experience building development tools that made it easier to collaborate and widely distribute software. They could simply put their code online and let word of mouth and network effects do most of the rest. It took a long time—with lots of bitter fights and lawsuits along the way—but eventually, open source became the rule rather than the exception.
Google led the corporate charge in the early 2000s. Instead of buying expensive operating systems, Google ran Linux on the servers in its data centers. Then, it took open source databases and file systems and wrote its own open source applications to fill in the gaps. This reliance on free software made it easier for Google to afford to give away services such as search, email, maps, and others. Facebook, Uber, Netflix, and many others would do the same. Today, open source is the engine of most major computing advances. Amazon.com Inc.’s massive cloud networks rely on Linux and many other free apps to function, which means that the tens of thousands of businesses that buy computing power from Amazon’s data centers are living the open source lifestyle, too. Google has placed Android, a variant of Linux, on more than 1 billion smartphones.
Thousands of people have contributed to Linux’s position at the heart of everything from TVs to cars, but Torvalds, now 49 and living in Oregon, remainsfirst among equals. In addition to writing the core Linux code that undergirds the internet and smartphones, Torvalds wrote a program called Git over a weekend in 2005 to help him manage the development of Linux. Some open source coders took Git and turned it into GitHub.
Torvalds draws a healthy salary from the Linux Foundation, a nonprofit funded by companies such as Google, IBM, Huawei Technologies, Tencent Holdings, and Intel to further develop the operating system. His total annual compensation of about $1.8 million is more than enough for him to buy a nice house in Portland and do as he pleases, which mostly means sitting at home coding. But if he’d been more interested in financial rewards and the daily grind, the guy might well have Bill Gates money. On paper, the company that’s made the most money from Linux is Red Hat Inc., which has created a custom version of the operating system and charges client businesses to keep it updated and secure. IBM acquired Red Hat for $34 billion earlier this year in the biggest-ever software deal. GitHub users can also opt to sponsor coders or projects that interest them, à la Kickstarter or other crowdfunding sites. Often, though, open source coders don’t get paid what they’re worth, and their status as hobbyists complicates the corporate world’s reliance on their work.
About this time last year a 48-year-old software developer in Sweden named Daniel Stenberg received a panicked call one evening from a large German automaker. The car company, which Stenberg declines to name, asked that he fly to Germany immediately because an application Stenberg had written was causing the entertainment system software in 7 million cars to crash. “I had to inform them that, you know, this is a spare-time project for me and that I have a full-time job and can’t just go to Germany for them,” Stenberg says. “They started out pretty demanding, but then switched when they realized the situation they were in.”
This is fairly typical for Stenberg, who since 1998 has been refining a widely used open source tool known as curl. Over the years, curl has found its way into the electronics of almost every new car, as well as software written by the likes of Apple, Instagram, YouTube, and Spotify Technology. On any given day, more than 1 billion people will unknowingly use curl, which helps transfer data between internet-based services. Developers from major companies and startups alike have grabbed curl off GitHub and elsewhere and inserted it into their products in ways that Stenberg could never have field-tested himself, and they’re not shy to send him messages at all hours demanding that he fix bugs promptly.
“Most of the days … I tear my hair when fixing bugs, or I try to rephrase my emails to [not] sound old and bitter (even though I can very well be that) when I once again try to explain things to users who can be extremely unfriendly and whining,” Stenberg writes on his website. “I spend late evenings on curl when my wife and kids are asleep. I escape my family and rob them of my company to improve curl … alone in the dark (mostly) with my text editor and debugger.”
In similar fashion, thousands of labors of love have found their way into software running everything from cash registers to trains. Software tools like GitHub have made this process easier with each passing year. Rather than rewriting every piece of an app from scratch, a developer just searches the vast library of open source code to grab what already exists. The end result is a complex system of interdependencies on thousands of freely available tools and apps. If one of the volunteers responsible for maintaining and improving those tools and apps decides he’s had enough, entire swaths of the internet and our infrastructure can cease to function until someone else steps in with a fix. “It’s a bit crazy,” Stenberg says. “Open source is a huge part of everything now, and I think it’s still growing.”
This isn’t quite the future the hippies wanted. In the beginning, free software zealots were trying to democratize technology, not create a way for powerful corporations to get more power. They wanted to ensure the best computing tools and data wouldn’t be centralized and metered out by corporations. They wanted people to have the freedom to explore technology and ideas away from the watchful eyes of an overlord.
Yeah, well, oops. Google, Facebook, Amazon, and many others have used open source code to create grand, global advertising networks that track and analyze billions of people’s every move, online and off. By comparison, ideological wars about bundling Excel and Internet Explorer with Windows 95 seem downright quaint. “If you don’t have control over the technology that runs your life, the devices and services that run your life, then your life will be run by other people using the computers,” says Eben Moglen, a law professor at Columbia who’s spent decades at the fore of the free software movement. “We made good stuff, and it was turned into ammunition against our dreams.”
Moglen says he appreciates the leveling effect that GitHub can have—it’s one of the best places for a talented 16-year-old programmer in Cambodia or Nigeria to show off her skills and alter the economic course of her life. Still, Moglen is counting on young people to form the core of a greater backlash against big tech companies’ privacy grabs. He’s pitching a hardware-software package called the FreedomBox, which costs about $90. It’s a small computer that uses open source software to replicate many of the common internet services (search, messaging, file-sharing) away from the prying eyes of the tech giants.
Other open source veterans argue that the revolution was worth it. Small teams of scientists can now punch well above their weight thanks to GitHub shortcuts. Cancer researchers, to cite one of many, many examples, frequently borrow from Google’s open source machine-learning work in their hunt for better ways to screen for tumors. “I don’t know who is religious about open source anymore,” says Dave Rosenberg, a veteran software executive and investor. “I don’t think you can achieve the stuff we want without it.”
Friedman, who spent 20 years starting open source companies and working on similar projects at large software makers before Microsoft put him in charge of GitHub, has the future of open source very much on his mind a couple days before the Svalbard trip, in Oxford, England. No true prepper is content with only one backup plan: The Arctic cave is just the first of what GitHub plans to be many repositories of code scattered around the world, holding almost all the code in its data banks rather than just the favorites. At one stop, Friedman climbs a few flights of creaky stairs to visit the head of the Bodleian Library, which keeps 12 million items in its glorious medieval towers. Would Oxford also store some code for safekeeping? As it turns out, sure, they’re game. Torvalds and Shakespeare, together forever.
In the spirit of the Svalbard cave, Friedman’s immediate mission is to tame the existential risks facing open source software. During our time together, he recounts story after story of large companies that have no idea how much open source software they depend on, who wrote it, how old it is, or what security holes might exist in it. He’s hoping that Semmle Ltd., a security research company GitHub recently acquired, can help close those gaps. GitHub is also refining the parts of its user interface that show a business what code it’s using, where that code is from, and when it needs updating. Yet another important step will be the creation of a more formal system for uniting big companies to subsidize volunteer efforts like curl, he says. There should be an easy way that Apple, Spotify, and the unnamed large German automaker can split the cost of a meaningful full-time wage for Stenberg with a few clicks.
“We would be successful if we could create a new middle class of open source developers,” Friedman says. “If you do this right, you create more innovation.”
GitHub’s most existential mission feels more urgent a few hours after we leave the Svalbard code cave. Fires have broken out around Friedman’s family home in Sonoma, Calif., and his wife calls to say that she and their 3-year-old daughter are evacuating. Friedman tells his wife to turn on their Tesla’s Bioweapon Defense Mode, which filters the outside air in extreme fashion. By the time dinner rolls around, he knows his house has been reduced to mostly ash. A photo of his front door archway—all that’s left standing amid the smoking rubble—soon becomes the image most media outlets choose for their coverage. With much of California burning or blacked out, an Arctic reset button starts to make a lot more sense. As Friedman has said several times by now on our trip, “I think the world is fundamentally weirder than it was 20 years ago.”
To contact the editor responsible for this story: Jeff Muskus at email@example.com, Jim Aley
©2019 Bloomberg L.P.