Can Tech Workers Take on the Oligarchs?

In 2022, JS Tan and Clarissa Redwine, both tech workers turned union organizers, were working 14-hour days trying to organize a strike at a company disappearing around them. Elon Musk had bought Twitter (now known as X) and was slashing the workforce. The two organizers called every Twitter employee they could still reach, but when the cuts were finished, just one member of the organizing committee was left.

Tan and Redwine’s new book Against Tech Oligarchy is not another lengthy case against the tech authoritarians so much as it is a counterhistory of worker rebellions from below. The authors argue that tech bosses became radicalized in response to worker organizing inside the industry—that organized workers are what inflamed the bosses’ worst fears of democratic accountability to unchecked corporate power.

According to their book, Tan and Redwine had gone into the fight at Twitter with the “recognition playbook:” build majority support for a union, win a vote to have it recognized by the National Labor Relations Board, then bargain a first contract—a process that can take years due to employer intransigence and lawbreaking. Their book now argues that the recognition playbook was a trap that deferred class conflict—that it taught tech workers to put off the fight with their bosses at the moment they most needed the practice.

I spoke with them about what happened at Twitter, why strikes are difficult in tech, and what workers can win by acting like a union even before achieving recognition. The conversation has been edited for length and clarity.

I’ve been reading some of the recent books about the radicalization of the tech billionaires—Jacob Silverman’s Gilded Rage, Karen Hao’s Empire of AI. What stands out about your book is that you argue that radicalization happened partly in response to organizing by workers inside the industry. What was the impetus for writing it now, when workers are facing both an employer and a government offensive?

JS Tan: What we really try to show in the book is that there’s a dynamic between workers and bosses—a class struggle—and that’s part of how we understand the radicalization of the tech bosses.

If you were a tech worker in the 2010s, you were making good money, bought into the company mission, and showered with perks. Tech bosses portrayed the industry as one that wasn’t going to do harm—“Don’t be evil.” They were selling tech as a haven for progressive workers. So workers took them at their word and started organizing on that ground. But by 2022, fed up with workers telling them what to do, the bosses realized that to stop tech worker organizing they needed to put an end to the industry’s progressive culture, and so they set out to rewrite the industry’s DNA—clamping down on DEI, eliminating AI ethics groups. Zuckerberg talked about wanting to bring more “masculine” energy to the company.

Silicon Valley’s biggest bosses have as a result now cozied up to the political right and have formed a tech oligarchy. There is little in society to hold this group of people accountable, and we both believe that it is within their own workforce that there is the power to push back against them.

The conditions in the tech industry make the need for organizing more urgent than ever, especially with AI, which is already being used as a pretext for mass layoffs. At the same time, the tech worker movement that we have spent so much of our time in has hit a low. You can see it in the decline in the number of collective actions and new union drives across the industry. There are many reasons for this, but we felt that part of it is related to the strategies tech workers and bosses have used against each other. We wanted to write this book to reflect on these strategies and ultimately to set tech workers on a new path to finding leverage.

Take me back to those early years. What did workers win, and what happened afterward?

Clarissa Redwine: At the beginning of the tech worker movement, tech workers put real pressure on their bosses and won huge gains: Google agreeing not to renew Project Maven [a Pentagon contract to build AI for analyzing drone footage] in 2018; Amazon agreeing to a climate pledge in 2019. Within a few years, though, Google was back in talks with the Pentagon, and Amazon had quietly dropped the central pillar of that pledge.

Tan: So we flipped 180. We felt like we didn’t know anything, and that we needed to take templates from other industries and from the past. That’s why we moved so hard into the union playbook.

Redwine: When we shifted into building unions as a movement, we were hoping to build more sustained pressure that would outlive the cycle of one news story. Years went by where tech workers shifted their energy to building toward formal union recognition by employers. We thought that the dip in public collective actions would have a payoff at the end of that road. Not only did we not see a payoff, but we entered a far more antagonistic climate from bosses.

No better story illustrates these shifting dynamics than your effort to run a traditional union recognition playbook to organize Twitter’s roughly 6,000 workers, as Elon Musk took over the site in 2022. If the union had succeeded, it would have become one of the largest tech unions in the world—but the effort fell apart. What happened?

Tan: Some context first. For the preceding decade or so, interest rates were near zero, demand for tech workers was booming, and there weren’t enough of us. We had incredible labor market power, incredible leverage. We were highly autonomous. Then in 2022 the Fed raised rates dramatically and tech companies started cutting costs—and tech companies are asset-light, so their biggest cost is their employees. Tech workers lost labor power almost overnight. That’s the backdrop to the Twitter takeover.

From the inside, we realized that Musk’s objectives were not just about cutting costs, but also very much about disciplining what he saw as the liberal havens within the company. So off the bat, he fired the AI ethics team; he canceled the human rights team and completely restructured content moderation. Ultimately, he cut around 80 percent of the workforce. Much of that was self-selected: he gave everyone 26 hours to commit to being “extremely hardcore” or take three months’ severance, and the people most likely to take that deal were the ones who’d been organizing. The ones who stayed often had no choice—a lot of them were on H-1B visas, and with the job market collapsing, nobody could afford to be out of work. After the next round of layoffs, there was only one member of our organizing committee left.

Redwine: It wasn’t just an attack, but a celebration of dominance. Musk turned Twitter itself into a stage showcasing how he was disciplining his workforce, where he called his workers lazy, undeserving, and woke. He found in a closet a bunch of “Stay Woke” shirts made by Blackbirds [an identity-based employee resource group at the company] and paraded them across his platform—made a show of denigrating these workers. That was permission for bosses all across the industry to not only start disciplining their workers and to clamp down on tech’s progressive culture, but to justify it to the public, which was something we hadn’t seen before.

JS and I believe that the organizing tactics that we chose at Twitter might have worked well under different circumstances, but when we were in that kind of environment where the boss was hellbent on busting a union and enacting mass discipline, those tactics just were not strong enough to move us to what we needed for collective action.

You both describe how easy it was for tech workers, especially in those years of the hiring spree, to switch jobs. In organizing, we talk about individual versus collective solutions. If you can just leave, that’s a very individual solution to a collective problem in the workplace.

And then there’s something about the workplace itself that becomes a challenge. Many of the workplaces that you’re describing have small teams of workers. They are not like the factories of the 1930s, which aggregated thousands of workers in one location. They are these sprawling nodes across different geographies and languages. So how do you build class consciousness in such a segmented workforce?

Redwine: In the tech workplaces that JS and I were trying to organize, there were several things that were particularly challenging. A company’s mission can feel unifying across class divisions, which makes it very difficult for workers to see their bosses as holding a different position from them. Also, the way that tech work is structured, almost everyone on a project team may have a different role, so it’s more difficult to figure out what would be the unifying issue across the people that you have the strongest relationships with.

Something else that we ignored for so long as we were building toward union recognition was that almost no tech worker had ever experienced a union before. Tech workers understand the concept of taking collective action, but the idea of going straight to a union is much more difficult, and a longer road, because of the unknowns.

Tech workers throughout the movement have come to class consciousness when they actually fight their boss for material change—whether that’s to uphold the company’s mission or to make the workplace better or to protect society. That is a critical moment where they understand the dynamics of the workplace, that they are on one side and the boss is on the other.

Last summer, I spoke to the Alphabet Workers Union after they won voluntary exit packages, covering about 60,000 workers at Google. And one of the things that struck me is that for many years, as a minority union [one that hasn’t won a workplace vote to be recognized by the National Labor Relations Board] they had hit a ceiling in terms of organizing around different political causes. But once they started talking about job security, some of their coworkers who were on H-1B visas and other workers who were worried about getting laid off started to come into the union, so their ranks grew.

I’m curious how that stands up against what I hear in the book, which is that you want to recapture dynamism of the early movement, the organizing around sexual harassment, the organizing around military contracts. And then there’s this larger question about how we build a durable organization.

Redwine: Something to keep in mind is we’re not necessarily advocating for any particular type of issue campaign. What we’re trying to clarify in the book is there’s a difference between building an issue campaign to win material change, and an issue campaign that is designed to get union recognition.

At the beginning of the tech worker movement, before there were the constraints of what a traditional union recognition drive looks like, workers were committed to building strategy that could actually put pressure on the boss, and that’s what public pressure did at that time. No longer the case—as companies have learned to absorb the bad press—but at the time that was effective.

What JS and I trace throughout the book is that when tech workers moved away from direct actions where they were clashing with the boss to win something material, and instead tried to build toward union recognition, they lost that experience of having class conflict. If you grow without ever having thrown a fist, you’re not going to be an effective fighting union.

Let’s talk about leverage and the strike weapon in tech.

Tan: The bosses in a lot of workplaces know that as soon as you have the strike threat and as soon as workers walk out, revenue stops. But for tech, these assumptions have undermined the movement in some ways. For tech workers, the fundamental part of their job is you’re building these products that then go on to live without you, which can effectively automate the revenue and the profits of your employer. So when you go on strike, the platforms will continue running, the ads will continue being served, the profits will continue being generated.

That’s what we’re trying to get at with this book: tech workers, at this point, haven’t figured out where their leverage is. We haven’t figured out what is the most effective strike, and I don’t think there is one answer for all tech workers.

Redwine: In the weeks before the Twitter layoffs, when we started circulating a strike pledge, the first thing Musk did when he saw it was freeze the codebase, and that changed everything.

A version of this happened at The New York Times. Workers timed their strike to election coverage, but management had them working 60-hour weeks beforehand to gold-plate the platform, and had them write documentation for contractors who were brought in supposedly to help—scabs in advance. By the time they walked out, they’d built the thing that let the Times run without them. [Editor’s note: The company said at the time that it respected the union's right to strike but called the timing "unnecessary and at odds with our mission."]

Only when you practice striking and actually try to bring the boss to the table do you learn how they act in that moment for next time.

You say that workers must engage in class conflict instead of deferring it for some big showdown. Many labor organizers would agree with you. And for decades labor scholars like Kate Bronfenbrenner and others have argued for escalating pressure tactics in union organizing campaigns, as well as acting like a union before winning union certification or bargaining a contract. What can tech workers do before they have the numbers?

Redwine: Class conflict does not have to start with every single person in the workplace; dozens of workers can win something material, not just for themselves but for the working class at these big tech companies. And then when they expand their solidarity to other parts of the company, they are seasoned organizers at that point. They’re not speaking from theory; they’re speaking from experience, and that is what we need to see more of.

Tan: We use the metaphor of piling tinder for a bonfire. We can’t just be taking all the tinder and never even presenting what striking a match would look like.

Redwine: I think one thing that brings me a lot of hope is that it’s not just us who are seeing this need for a shift. At Circuit Breakers, a labor conference in tech, we were hearing workers who were coming into the movement fresh, who had this perspective already of, “The union’s going to take a long time. Let’s start acting like a union now and start winning what we need now because there’s no time, y’all.”

There was a real win that came out of this shift not too long ago. I joined a campaign at Meta where the workers were attempting to push back on some pretty egregious automation from the boss. The workers could have done what we had done in the past, where they say, “Hey, worker, you really care about not being automated by AI or not having your data scraped? Join the union. Let’s build toward majority support, and then we’ll fight for issues like this.”

Instead, the workers said, “Every person who comes into this campaign, we are going to build a strategy to win dismantling of this initiative as soon as possible,” and they did win, and that is the shift that we’re hoping to see all across the industry.