Our new issue is out now. Holiday subs start at $5 if you follow this link.

Microsoft and the Yeoman Coders

Microsoft’s purchase of Github is the latest chapter in capitalism’s oldest story: the absorption of artisan labor into the circuits of capital.

The Microsoft store on Fifth Avenue in Midtown Manhattan is shown June 4, 2018 in New York City. Drew Angerer / Getty

This past week, Microsoft acquired open-source code repository GitHub. In the longue durée of the tech industry, it was a world-historical about-face: open-source code, which allows other programmers and interested parties to view, and even copy and modify code, arose in opposition to software companies that rely on strong intellectual property protections to wall off their products from technologically skilled users.

In its early days, personal computing was a hobbyist pursuit in which young men (and it was mostly men) tinkered collectively with mail-order kits. Commercial software was scarce: you had to program things yourself. Within this milieu arose a robust culture of cracking things open and sharing what insights you found. Learning from others’ code was commonplace, as was passing around programs.

As we know now, one person’s “sharing” is a property owner’s piracy. Instead of buying programs, hackers were copying them and making their own. In 1976, a young Bill Gates, trying to get his software company off the ground, sent a now-infamous letter to the Homebrew Computing Club newsletter that drew a line in the sand. “As the majority of hobbyists must be aware, most of you steal your software. Hardware must be paid for, but software is something to share. Who cares if the people who worked on it get paid?” Beyond snatching dollars out of the pockets of the soon-to-be richest man in the world, piracy created a problem for software itself. “One thing you do is prevent good software from being written,” Gates wrote, before helpfully offering a mailing address to which software pirates could send their belated payments.

But was it true that sharing undermined the quality of software? Not everyone agreed. In fact, many programmers, who had learned their craft by looking at how other programs worked, believed the opposite: sharing code meant more people became better programmers, giving them vital skills in a world where computers were quickly taking over everything. Furthermore, if more people could tinker with code, there were more potential mechanics who could fix errors: “with enough eyes, all bugs are shallow,” a later slogan would put it.

But the law was on Gates’s side. In a somewhat bizarre ruling, the Supreme Court held that a piece of computer code is an expressive work, like a poem or a story, rather than a mechanical invention. This meant that programs would be covered by copyright law rather than patents. While patents enter the public domain after twenty years, copyright lasts much longer — and, due to aggressive lobbying by entertainment corporations, the length gets longer and longer.

Enter Richard Stallman, a computer scientist ensconced at MIT. Insulated from the vagaries of the market by academia (unlike Gates the entrepreneur), Stallman sought to protect the guild-like practice of skill-sharing from which he and his associates learned to program. In a celebrated hacking of intellectual property law, Stallman devised alternative licenses, part of a “copyleft” initiative, that allowed programmers to designate their code as free to share. Further, the GNU public license declared that all software developed from this free code would itself have to be free. This meant that free software would spread like a virus. As Stallman put it, “Copyleft uses copyright law, but flips it over to serve the opposite of its usual purpose: instead of a means of privatizing software, it becomes a means of keeping software free.” Or in the words of former Microsoft CEO Steve Ballmer: “a cancer that attaches itself in an intellectual property sense to everything it touches.”

Stallman’s work captured the hacker zeitgeist, and his polemical writing helped spread a hearty disrespect for copyright that permeated digital culture for decades. But free software wouldn’t have its true proof of concept until 1991, when Linus Torvalds, a computer science student at the University of Helsinki, put it to work for the operating system he was developing. Not only would it be a free and open competitor to Microsoft’s hated Windows, which then appeared on the brink of capturing the entire personal-computing market; it was also the first major example of successful crowdsourcing. Torvalds leveraged a passionate global community of coders to collaborate on the program — dubbed Linux — as much or as little as they liked.

Linux was always too technical for the casual computer user, its appeal confined largely to those already steeped in hacker practices and values, such as an abiding antipathy to Microsoft. But Linux was greeted among the digerati as the latest example of how the power of networked computers was changing everything for the better by empowering individuals — if only that troublesome copyright beast could be tamed.

Economist Yochai Benkler viewed free software as an example of an emerging mode of production that disregarded both intellectual property and wage relationships. This “peer production” took place in a “commons,” rather than a market, where people could engage in non-alienated work and still produce marvels. Ethnographer Chris Kelty theorized this development process as a “recursive public of geeks” who crafted a “moral-technical order” through debate and discussion: open source had fashioned a hacker public sphere. Anthropologist E. Gabriella Coleman, who studied the Debian GNU/Linux community before she moved on to the hacker group Anonymous, cited a hacker-crafted haiku that spelled out how to make a program that would crack DVDs. Code is poetry after all! Code is speech!

Hackers may be creative and generous, but treating code as speech, and programmers as free speech partisans, also obscured some important elements. Software programs — even those hewing most closely to the holy precepts of the GNU Testament — were more like linen and coats than poems: they were valuable commodities. How does free software work within capitalism?

While there’s a lot of hype about gift economies and good vibes, when you’re doing an economic analysis, you have to consider the sector as a whole, not the intentions of a few members. One of the few to do this well, Swedish researcher Johan Söderberg, analyzed free software as part of labor under capitalism. Because so much code is available at no cost, open-source programmers effectively cheapen software by creating a base of code freely available to all — in Marxist language, they’ve cheapened those commodities by reducing the socially necessary labor time required to create them. While Stallman’s copyleft licenses envision a society of individual developers and users all benefiting from accessible and customizable software, there’s nothing that prevents corporations from getting in on the action. In fact, entrepreneur Tim O’Reilly, recognizing the value of free software, rebranded it by replacing that ambiguous word “free” with “open source,” to signal that it had no inherent antipathy to profit-making.

Subsumed Into the System

Today, the biggest companies, like Facebook and Uber, rely on bits of code snatched from open-source public repositories, and the boom in app startups over the past ten years was driven by the abundance of good free code, making an app easier and cheaper to launch. You don’t need a stable of coding geniuses, just people with enough know-how to pick the right bits of code and stitch them together. Open source is so valuable that its projects even attract venture-capital money without clear prospects for direct monetary returns. The ecosystem itself provides value. As Microsoft’s current CEO Satya Nadella put it, “We are all in on open source.”

Even if you’ve never struggled with a Linux build on your own laptop, free and open-source software affects you. It makes up a huge part of the digital infrastructure, from routers to web browsers, developed by everyone from Walmart to ExxonMobil to the International Space Station. Even Linux, nominally free, has spawned large firms, like the multibillion-dollar service company Red Hat.

Although free-software developers generally have little love for corporate behemoths, their organizational forms proved perfectly amenable to digital capitalism. Stallman may have rendered his manifesto in that very American language of the defense of freedoms, but what he was really defending was his own independent and collaborative working conditions, which included disregarding copyright as he saw fit. Rather than protect civil liberties, free and open-source software is best understood as a kind of craft-worker struggle, about preserving autonomy and skill.

For example, according to a recent report by the Ford Foundation, the success of open-source software has given a lot of mobility to developers as a class. Free and open-source software popularized nonproprietary coding languages, making them the default for most technology development, and thus granting programmers a high degree of flexibility in their work. If you know Python, that skill is useful at many different companies. Don’t like your job? Quit and go somewhere else.

This helps to explain why many hackers have understood the problem of corporations as one of “centralization,” rather than of capitalism as such. When they face problems at work, the default strategy is often simply to jump ship — or better yet, build an alternative themselves. And “decentralization” is often baked right into those alternatives. Free-software social networks, like Diaspora and Mastodon, trumpet their decentralized architecture, which, according to Mastodon “avoids the risks of a single company monopolizing your communication.” And of course, Bitcoin and its underlying technology, the blockchain, are the latest hacker innovations that fly the flag of no central authority.

The belief that things are better when they’re run by small independent producers goes to the heart of what Andy Cameron and Richard Barbrook, in their classic critique of Silicon Valley “The Californian Ideology,” diagnose as the “Jeffersonian” nature of tech culture, predicated on independent small production. Every coder is a digital yeoman farmer, preserving an independence grounded in individual abilities. Cameron and Barbrook align this with a libertarian ethos, but, as I argue in a forthcoming paper, a case can be made that hacker culture can also be understood as Proudhonian, drawing from the same well of concerns as mutualist anarchist artisans of the nineteenth century.

But decentralization has its limits and contradictions. First, it’s tough to manage large software projects in a decentralized manner. This is the problem GitHub solves. It owes its existence to its elegant solution for managing software projects, as programmer and writer Paul Ford details in Bloomberg. Without getting into the weeds, GitHub smooths the chaos of a dispersed set of developers working on the same programs. People use it, and congregate to it, because it works.

Second is the so-called “network effect.” GitHub is not just a place where people put their code, but also a social network of sorts. Programmers build their reputations through their record of contributions to projects. This leads to jobs. Two decades ago, Benkler marveled that so much of the work of coding was done unpaid and voluntarily. Now we understand better: developers earn social capital that can translate into real money down the line. So, GitHub, in effect, is part of a centralized world market for coding labor — 28 million users and 57 million repositories of code. Several commentators have drawn a link between Microsoft’s acquisition of GitHub and its recent purchase of LinkedIn: it wants to corner the market in “talent.”

Finally, centralization is not simply a design choice. It is what Marx called an immanent law of capitalist production. In a competitive phase of capitalism, many firms compete. But the rigors of competition mean that firms face pressure to mimic successful strategies (for example, acquiring massive amounts of user data to sell to advertisers, which essentially all commercial platforms, no matter their purpose, do). Further, competition means that many businesses will lose or give up, and thus cede market share to the victors. “One capitalist always kills many,” Marx wryly notes, and we could say the same for social networks and search engines. GitHub’s fate is the same as practically any successful startup now: it quickly gets snapped up by one of a handful of massive corporations. If it hadn’t been Microsoft, it might have been Google instead. If your project exists within capitalism, it will face these structural issues, no matter your good intentions.

When I first got online, it was common knowledge, spread by free-software partisans, that Microsoft was the enemy of everything good, and that only eternal vigilance and GNU public licenses would keep Big Brother Bill Gates at bay. Now the fox is in the henhouse, and needless to say, hackers are nervous. In response, some programmers moved their projects over to rival hubs like Gitlab, but that’s a bit like quitting Twitter to go to an off-brand network none of your friends use. GitHub, benefiting from network effects and capitalism’s inherent tendency towards centralization, is likely to remain on top.

Historically, programmers have maintained a large degree of autonomy in their work, along with high wages. This happened for conjunctural reasons, unlikely to be replicated (a series of boom-and-bust cycles of overproduction), as well as the unique forms of organization and struggle they waged against intellectual property and corporate control. It afforded coders independence and the freedom to take their business elsewhere if working conditions weren’t to their liking, and even to build their own autonomous zones, however temporary.

But this situation is rapidly changing, and there are fewer escape routes than ever. Desperate to drive down the wages of its expensive employees, Silicon Valley has launched initiative after initiative to increase the reserve army of labor by teaching people to code as quickly and cheaply as possible. Low-cost advanced IT work in English-speaking areas overseas, such as India, has already become a crucial part of the industry. Monopsony power over the labor market will reduce coder autonomy, and tech companies have outright colluded to suppress wages in the recent past. Innovative licenses won’t help — indeed, a new generation of coders ignores them, in favor of just posting everything to GitHub.

The changing terrain of digital capitalism means the old strategies of coding freedom won’t work. So instead of attempting to decentralize tech, to better preserve an ecosystem of independent businesses that stay small and promise to be good, coders should strive for democracy at work instead. There are a number of encouraging signs that tech workers are moving away from the narrow craft sensibility and looking to mass worker organization for models. Union talk has been in the air for a while now, and recently a major conflict broke out over a serious union drive at logistics firm Lanetix. NYC-DSA’s Tech Action Working Group and the Tech Worker Coalitions in Seattle and San Francisco speak to this emerging perspective.

Further, a centrally organized and democratic tech sector might solve many of the significant problems with software that neither the voluntaristic open-source ethos or the profit-maximizing prerogatives can address. As the Ford Foundation report points out, coders and capitalists tend to be drawn to new and exciting projects, leaving upkeep of our existing open-source infrastructure, much like our physical infrastructure, to decay. With enough eyes, all bugs are shallow, but not if those eyes are drawn away from older projects to sexier ones. Our world needs maintenance as much as, if not more than, innovation, and a democratically organized tech sector could be the cleverest hack of our system yet.