What I Fought For

Before reading this, please read Jeff Vogel’s excellent essay, “How I Deal With Harassment, Abuse, and Crazies In General“. His experiences, as one who has encountered both the positive and negative aspects of publicity, are eerily similar to mine.

“A Reputation Problem”

• [I got some career advice from one of those people who finds CTOs at top hedge funds in NYC and Chicago, and he said, emphasis mine]: You’d be one of the best people they’ve seen in years, and unlike the people around them, you’re honest and ethical. The thing that’s holding you back is… a reputation problem. People Google you and think, “This guy’s going to start a union.”

Brass Tacks

• There’s a misconception about me, which is that I attempted to unionize Google. Nothing of the sort happened.
• Silicon Valley is aggressively opposed to unions. Companies routinely share lists of suspected unionists, and people who end up on these lists can be blacklisted, assigned to the worst managers if “discovered” once inside the company, harassed in public, and subjected to all sorts of libel.
• I’m not even “a unionist”. I just choose not to reject the idea out of hand.
• So what do I believe?
• First of all, I believe that collective bargaining is a right. Corporations are collectives that bargain on behalf of the rich people who own the means of production. The workers ought to have similar backing. It’s only fair. In fact, most respected professional organizations are, in fact, labor unions under a more genteel name.
• Second of all, I recognize that there is a diverse array of collective-bargaining arrangements, and that not all are good. […] The negative reputation of unions comes more from propaganda and clever redefinition of the term “union” than anything else.
• Traditional labor unions emerge (and, often, work very well) when the labor has been deemed by both sides to be a commodity: for example, coal production. If commoditization of the labor is irresistible, then the union tries to ensure that it happens on fair terms.
• In fact, the strongest objections against programmer unions come from software engineers who resist allowing their work to be viewed as a commodity. To this, I am sympathetic and in intellectual agreement. It would be a better world if we were artisans. That said, we’ve failed to oppose commoditization of our labor for as long as we’ve been without a collective arrangement[…]. If we believe that our work is a commodity, then we should not be resistant to unions. If we truly believe that it’s not, then we ought to consider a professional organization or a guild system that can defeat the commoditization that has happened this far.
• Professional organizations are unions that exist to uphold ethical principles […] that ought to supersede both economic concerns and managerial authority; they exist to dictate terms of partial commoditization. Likewise, high-end unions (as in Hollywood and professional sports) and guilds exist to enable those who view themselves as athletes or artists […] to coexist with commercial players.
• Third, I am (surprisingly?) more skeptical of software engineer unionization than supportive[…]. We’ve had the worst people driving out the good for at least 15 years, if not longer. We’ve ended up with an exclusionary macho-subordinate culture. We have sexism, ageism, classism, racism, and widespread harassment in our midst. We can’t blame all of this on management. If we create a union without fixing our culture at the same time, we’ll probably generate one of the bad kinds of unions (the garbage-in, garbage-out principle). We should fix our culture first. That’s the top priority.
• Fourth, we can nonetheless learn from existing professional and trade unions, and start figuring out what will and will not work. Most programmers believe that all unions negotiate (and regulate) compensation and that, therefore, top performers will see mediocre wages. That claim is demonstrably untrue[…]. Most likely, an effective union for software engineers would not try to set salaries, but would give employees additional rights when it comes to how they are managed. We could kill stack ranking and improve working conditions, and I don’t see a good reason not to do that. We could guarantee employees the right to an independent representative on issues surrounding performance management[…].
• Fifth, it’s important to note that it is impossible to unionize a workforce that doesn’t want to have a union.
• Let’s be honest here. Software engineers are individualistic even when it goes against their own interests.
• It’s simultaneously flattering and annoying that some employers have thought that I’m capable of single-handedly unionizing their workers, in addition to everything else that I have to do. In reality, that’s not a credible threat.
• Sixth, one should note that unions are often good for the companies that are unionized. The (incorrect) common viewpoint is that, since unions drive up workers’ wages, shareholders are losing money. This isn’t necessarily true. If the union’s effect is to make the internal labor market fairer and more efficient, and to improve workers’ conditions, that might actually result in the company getting more value out of its people.
• Who gets hurt by unions? Executives, the most. Managers lose authority and power, while their jobs become more complicated. They can no longer unilaterally terminate people, which means that extorting the workers into supporting their own career goals is no longer an option. For owners, however, unions often mean that a better product is produced. So, while wages increase, the net effect can be positive or negative for profits.
• Managers dislike unions because they’re perceived to be a vote of no confidence from their workers.
• Seventh, a company-specific union won’t work for software engineers. As in Hollywood, we have project-based careers and it’s normal to change companies frequently. The most important function of a technologist’s union will not be to protect jobs but to protect our reputations and our professional autonomy across the course of our careers.
• Eighth, unionbusting is one area in which otherwise competing malefactors will cooperate. This isn’t surprising, but its implications are disturbing.
• in Silicon Valley, there are employers who check references on people not provided (“back channel” reference calls). [T]hese companies want to share blacklists and bust potential unions.

The next 40 miles

• As of April 2016, there isn’t much evidence of a desire among software engineers to unionize, professionalize, or organize in any way that would threaten the status quo. This might be the case because the startup culture is powered by its low-level workers’ unrealistic expectations (e.g. the fresh college grad who works 90-hour weeks because he thinks he’s going to be a CEO in three years) and because those who have more experience are usually purged. The VC-supported startup culture of misogyny and age discrimination (see: open-plan offices) serves to exclude people with the diversity of experience that we’d need in order to organize, and I believe that this is the true purpose of that culture.
• The first thing that we need to do is fix software’s culture. We need to stop driving out women, programmers over 30, and people from non-traditional racial or career backgrounds.
• Why does Silicon Valley insist on hiring only young men? Because they don’t have the experience and organizational skill to threaten management’s interests. It’s easy to exploit them. While those fresh college graduates are almost never as good at programming as their more senior counterparts, that often doesn’t matter in a style-over-substance startup culture where companies exist to be sold quickly, rather than being built for the long term.
• I think that it’s important to solve the charlatan problem. We have a labor market flooded with incompetents, and we suffer under micromanagement frameworks like “Scrum” that are sold on the promise of making such people productive
• I think that the best way to drive out the non-programmers and the non-serious technologists is to implement an exam track similar to what the actuarial sciences use. We want to keep the programming profession open to everyone who has the ability, drive, and intellectual curiosity; while at the same time preventing management from flooding the labor market with an inferior-but-manipulable substitute: open-plan Scrum drones.
• What I like about an exam track is that it’s blind. Blind grading delivers a massive improvement over “culture fit”, under which a person can be rejected for not having played beer pong for ten years and not using PUA lingo in regular conversation.

She rides for Zelos

• as a community, we need to start thinking about the ethical ramifications of what we’re doing.
• There are the small failures that I see on a day-to-day basis. One trait of true professionals is that people inside the profession never criticize each other to outsiders. That’s not necessarily for the purpose of secrecy, but because the outsiders will lack appropriate context and become dangerous, even if they don’t mean to be.
• For example, let’s say that two engineers disagree about source code formatting.
  • What happens, though, when such a dispute gets escalated to non-technical management? To the manager’s ear, it sounds like one of the engineers is willfully refusing to conform to “correct” source code formatting and therefore singularly responsible for the disagreement. Whom will that manager judge to be in the wrong? Well, since there isn’t an objective right way to format code, the person with more political pull will win, and the other will lose and be labelled as the nonconformist.
• Non-technical managers exposed to engineering disagreements, but feeling a need to Make A Decision, often take those issues out of context and become TWiGs (Toddlers With Guns). As a general rule, I think that we won’t be respected as a profession until we develop the social and individual skills to handle these matters, whether they’re minuscule disagreements or genuine performance issues, internally.
• there are much bigger ethical failures that are even more worrying than our internal affairs.
• some industries clearly are actually evil. Employee time-tracking software is a major business. Who writes this code? Someone does.
• The question I’ve had to ask myself, again and again, is: does any of this stuff matter? When I’m reading math or CS papers at 9:30 on a Sunday night, what on earth am I working toward?
• I believe that we, as technologists, can take back our industry. To do so, we need to stop treating our careers as things that other people (managers) and institutions (employers) make happen to us, and step up and make things happen for ourselves. Toward this end, we need to work as a collective and develop a “slapping one is slapping all” mentality.
• On the same token, we’re in a state that is both degraded and paradoxical. In some ways, we need to make the software industry more inclusive. We need to halt the processes that drive out older programmers, women, minorities, people with disabilities, and others who are culturally vulnerable. We need to make a culture that includes and welcomes all people of talent, and not just people who look like me. At the same time, we need to make the industry more exclusive, in the sense of protecting everything that matters. There are a lot of programmers with no intellectual curiosity, nor any sense of craftsmanship, and those open-plan Scrum brogrammers (who implement the sexual harassment culture and the ageism that needlessly drive out talent) just need to go.
• Let’s talk about the near future. The VC-backed bubble is going to end, just like the one in the late 1990s did. The toxic unicorns are likely to die […]. However, my fear is that, due to our lack of organization, legitimate programmers might also see our salaries and conditions decline.
• What happens then? When jobs disappear in large numbers, worker leverage declines. Motivation to fix the problem increases (read: people get pissed off) but the ability to do so declines.
• The problem faced by those of us who are legitimate technologists is that the labor market has been flooded with an inferior product: the open-plan Scrum brogrammers who aren’t good or even acceptably mediocre programmers, but who look the part (“cultural fit”). They can’t fool real programmers, but they can fool non-technical management and even investors. They’ve also been able to sell, to management, a perception of flexibility. You can hire twenty-five of them in an afternoon, and they’ll work long hours and tolerate punishing work conditions. This would sound great, except for the fact that their technical abilities are abysmal. They’re so terrible, in fact, that they’re often negatively productive on a project with any technical meat. Worse yet, if they’re given any creative or ethical responsibilities, the results are catastrophic.
• Why has the market for programmers been flooded by inferior replacements? The perception is that technical excellence doesn’t matter, and that the market situation proves this. Unfortunately, short-term incentives often produce disalignment, and long-term consequences take so long to come to fruition that powerful individuals (such as technology executives) can avoid accountability for their own bad decisions.

Appendix: questions I get asked

“How is blacklisting enforced?”

Since these lists are illegal, they are usually only shared at an executive level.

The targeted person is usually subjected to negative, unsubstantiated rumors pertaining to health, work performance, and (until recently) sexual orientation. For every person in-the-know who engages in deliberate unionbusting, there are several “useful idiots” whose role is just to reliably repeat the gossip they are fed.

“Does Google still retain or participate in this sort of unionbusting?”

[I]f the company still uses stack ranking (and I do not know whether it does) then it is overwhelmingly likely that the unionbusting is just hidden in the stack-ranking machinery.

In fact, the point of employee stack ranking is, traditionally, to intimidate employees and prevent them from organizing. Rich companies actually don’t care about harmless low performers who draw salaries but do little or nothing.

“Why do you air ‘dirty laundry’ about ex-employers?”

I don’t. I made that mistake, once, with Google.

There are some who believe that “biting the hand” deserves a professional death penalty. Such people should be counted among the enemies of progress and treated as such. Nonetheless, I still believe that it is unwise, for an entirely different reason, to disparage an ex-employer.

That reason is this: in practice, the workers are more exposed to the corporate reputation than the executives (excluding the CEO). Bad-mouthing an ex-employer hurts the wrong people. That’s why I’ll almost certainly never do it again.

“How do you respond to people who believe that unions in Silicon Valley will kill its ability to innovate?”

Silicon Valley seems to be a nadir in terms of innovation, but the wrong labor structures could make it even worse.

However, a collective arrangement […] could also improve the situation greatly. How so? If people who actually make things get more respect, they’ll be more engaged and the products made will be better. If they make more money, they’ll be able to fund more interesting projects and “bootstrap”.

For the past twenty years, the biggest threat to Silicon Valley innovation hasn’t been some union bogeyman but the technology industry’s management. A common slur against unions is that they can promote a culture of complacent, arrogant mediocrity. That culture already exists. See: Scrum, open-plan offices, business-driven engineering. Unions could make that culture worse, but they could also drive it away.