A Google employee recently ended up in hot water for arguing that women and men tend to seek out different activities and careers because they have, on average, different values and interests.
This is an idea that–I exaggerate only slightly–every single person who has ever lived has probably entertained at some time.
So why did this guy’s memo provoke a firestorm of angry commentary?
We can talk about swelling resentment toward the tech industry, anger at the claims of evolutionary psychology, longstanding anxiety around sex and feminism. But having the read the memo, I think the writer made three critical decisions.
First, he wrapped his argument about sex differences in drawn-out attack on PC thought-policing, presenting himself as a victim of liberal groupthink. He picked a fight, and he got one.
Second, he used the trigger word “biology,” which always sets off rhetorical microwars between scientistic contrarians and postmodernist ideologues.
Third, he set up his discussion of sex and personality by asking what makes women different from men. No matter how you answer that question, liberals get angry, because it implies that men are the standard sex from which women represent a deviation. The memo writer might have had better luck flipping his argument. What, he should have asked, is different about men?
Consider what the reaction might have been like if he’d written something along these lines.
Google is admirably committed to hiring and advancing more women. Currently, the company pursues this goal through forms of diversity training meant to reduce discrimination. The hope is to make the workplace more welcoming by changing employee attitudes and behavior.
But what if that’s not enough? Are there other methods that might help our company–and the tech industry in general–recruit and retain more talented women?
That’s a question worth asking. Mandatory diversity training has documented drawbacks. It can result in superficial compliance while provoking a backlash that results in more discrimination behind the scenes. It puts the onus for change on individual employees, without addressing institutional factors that might be driving women out of the field. It draws inspiration from scientific research that has recently come under fire (see, for example, the literature on stereotype threat and implicit bias). And it focuses only on one half of the problem. Instead of asking, “What drives women away from computer science?” we at Google might want to ask, “What can help to draw them in?”
Here, too, current research can offer guidance. Tech is famously a male-dominated field. And the psychological literature reports that men are, on average, more likely to exhibit certain traits, preferences, and handicaps.
Many of these will sound depressingly familiar. Men have been found to be less cooperative. Men are more aggressive, and, as a result, perhaps, more willing to put up with stress and pain. Studies have found that men are less agreeable, less gregarious, and less empathetic.
Significantly, men have also been found, on average, to show less interest in other people. Sometimes this is described as a preference for things over feelings, sometimes as a taste for systems over relationships. Whatever the terms, the trend is the same. Overall, men put a lower value on their interactions with others.
None of this is true for all men. The average differences are often small, and individuals show wide variation. And it’s not entirely clear where these traits come from. Some have been linked to high levels of prenatal testosterone. Evolutionary psychology suggests that, in an unkind phrase, men are simply more “disposable” from a genetic perspective, therefore more likely to engage in risky and solitary behaviors. It goes without saying that culture plays a role; boys are often encouraged to pursue status at all costs and punished for exhibiting stereotypically feminine traits.
What can be said with certainty is this. At the age when young people are contemplating a choice of profession, men are more likely to consider jobs that are stressful, combative, abstract, and lonely.
That sounds a lot like the tech industry.
So what can be done?
We should all be cognizant of the baleful effect that discrimination has on workplace culture. But we should also entertain the possibility that structural and systemic factors help to make tech a male-centric field.
Instead of putting the blame for our lopsided labor force entirely on employee attitudes, here are steps Google might take to attract a more diverse pool of candidates.
- Make software engineering more people-oriented with pair programming and collaboration. Find more outlets and uses for employees who value cooperation. The tech industry in general should make a greater effort to attract talented candidates who thrive on social connections.
- Offer more stress reduction courses and similar benefits, and work to make leadership positions less stressful. The tech industry is famous for its punishing demands; this drives away sensitive but capable recruits.
- Offer more flexible hours and more opportunities for part-time work, making accommodations for employees who want a reasonable work/life balance.
- Focus critical attention on male gender roles, challenging cultural influences that drive men to seek status at the expense of human connection.
In pursuing these goals, we should be mindful of the company’s larger mission. It makes sense for competitive, hardworking employees to get ahead at Google, whether they’re women or men. As always, we have to balance the benefits of greater diversity with the costs of changing the way we do business.
But isn’t change what we do? In striving for greater diversity, we can’t allow ourselves to lapse into a dogmatic insistence on what’s always been done. We should experiment with new approaches, new views, new sources of information. We should let science guide and inspire us. Above all, we should entertain a diversity of views in our pursuit of a diverse workplace–and, as always, keep an open mind.
If he’d written that memo, people might have objected to its gender essentialism. Critics would have quibbled with its facts, assailed its logic. Male rights advocates might have griped about its depiction of loutish undersocialized men. But would furious debates have blazed across the internet? I doubt it. And yet the substance of the argument is largely the same.
There’s a lesson here about the ways in which different groups compete for victim status. But I’m more interested in the tech angle. As I’ve written many times, the web is something of a giant gossip machine, and nothing illustrates this better than our endless online tiffs and spats over some person’s rhetorical choices. There are probably millions of extant documents saying that men tend to be less empathetic, or that women are more likely to put a high value on intimacy, or that women are prone to beat themselves up over trivial matters, or whatever. God knows everyone talks about this in private life. But it doesn’t light up the circuits of cyberspace.
Then some day, somehow, someone pushes all the right buttons, and the gossip machine goes into effect, sputtering, churning, shutting down servers, doxing names, spitting out takes and takedowns and downvotes, elevating some reputations, ruining others, doing its work as loudly and efficiently as a woodchipper. Pop, fizzle, like, tweet, ding. At the end of the process, a few more ads have been clicked, a few gigabytes of data have been harvested, a few new ulcers have formed, and a few more citizens have been politically radicalized.
Does it have to be this way? Does a worldwide network have to devote some share of its bandwidth to running a virtual gossip machine? Or is this a function of the way we built out the software layer of the internet, going back to Google’s original use of links as an index of popularity, popularity as an index of relevance, and relevance as an index of informativeness? Almost everything about the current design of the web–everything about the way it organizes information–is based on a few crude assumptions:
- Something popular is preferable to something unpopular.
- Something current is preferable to something old.
- Something fast is preferable to something slow.
- Something familiar is preferable to something unfamiliar.
These masquerade as rules for sorting information well. But they’re really just shortcuts for sorting lots of information quickly.
I don’t deny that as shorcuts, they do a good job of approximating human decision making. If I’m presented with an overwhelming range of choices–a mountain of unsorted books, a bevy of similar newspaper stories–I’m more likely to choose something popular, new, short, and familiar. But why should I want my web tools to emulate me at my most mercurial, scatterbrained, hasty, and impatient? I want tools that help me make better decisions, not tools that always play to my worst tendencies.
It’s bad enough that the web works this way with content sorting, product recommendations, site rankings, and news. But when it comes to social connections–reputation, relationships, group affiliations, any kind of conversation–this quick-and-dirty approach is a disaster. The gossip machine takes in the signal of human intelligence and spits out the noise of human stupidity. It brute-sorts people by the roughest metric available, asking of every user: what drives you crazy? It’s like a piano that only makes music when you bang the keys with a sledgehammer. It’s like a bathroom scale that thinks you weigh more when you’re shouting. It’s like a superpowered calculator that only calculates common denominators.
Worse, it’s a machine that treats its own design as input. The system’s built to favor whatever’s recent, popular, and familiar. That’s what users see, so that’s what they respond to. Whatever’s popular gets more popular, whatever’s familiar becomes overfamiliar, and the timeframe for what qualifies as recent narrows to a razor-thin margin. The Google bosses like to imagine that the web will one day evolve into a vast AI. But it’s already an AI, thinking one idiot thought over and over: “Wow, people sure love trivial shit!”
I really believe we’re living in a dismal age for information technology. Not a dark age, exactly–it’s not like we’ve forgotten how to compute. A white-hot age, perhaps, an age illuminated by the light of ten billion digital suns, until our eyes are fried and the atmosphere ignites. We’re over the early flirtations, past the engagement, through with the honeymoon, recovering from the warm glow of infatuation. Now the house is full of unwashed laundry, there are dirty diapers under the bed, something’s burning in the oven, and the sink just broke. Someone you used to love is yelling that it’s all your fault, and the air smells like despair and charred shit.
It’ll get worse, I think, before it gets better. Like that part of the agricultural revolution when Egyptian peasants and slaves were dying by the thousands to build temples and tombs for crazy despots who happened to have a monopoly on literacy. Like that part of the industrial revolution when troops of workers were marching into the mouths of Satanic mills and children’s corpses were piling up in Irish workhouses.
We now have proofs of concept for the dream devices of science fiction–a worldwide network, virtual worlds, portable computers in every pocket–but the tools themselves are worse than bad, because they all rely on cheap fixes for major challenges. We’ve sorted the world’s information, but we used the crudest possible methods. We have computers in our pockets, but the interface is so constrained and clunky that all you can do is poke at them like an impatient toddler. We have virtual worlds that look increasingly lovely, but we get around them using clunky overlays that constrict attention and stupefy the senses. We traded our privacy for a giant pile of data, and we have no idea what to do with it.
Almost everything about the way we use computers today comes down to a simple trick: making computers better by making people worse. We haven’t mastered the nuances of human behavior, so we reached for the easy money. We staked everything we had on addiction, anxiety, distractability, and anger. We can’t even come close to emulating knowledge, but we’ve invented ten gazillion workarounds.
The crazy thing is that I think engineers know this. They know they’ve built a bloated, teetering cybereconomy on a clumsy hack of the human pleasure center. They know they’ve created an enormous complex of flashy but inflexible devices that can only function if users are trained to repeat a few compulsive and restricted actions. They know that those two features will only work if they work together, that the world of popular software is mostly a pile of Skinner boxes, training users to keep pounding at a few simple levers because the system would fall apart if they tried to do anything else.
They know, above all, that they’ve tapped into a tiny, tiny part of what computers or humans can do.
They know this is their world, and it drives them crazy, which is why they’re so frantic about taking the money they’ve made off their gossip machine and pouring it into moonshot projects, AI research, robotics, and space exploration. When I look at the Google bosses or Mark Zuckerberg, I see a bunch of aging pop stars using the money they made licensing bubblegum hits for car commercials to self-fund an experimental jazz career. “Look,” they’re telling the world. “Tech doesn’t have to be top-forty all the time. Computer science can be so much more.”
But the public isn’t listening. The public is stroking its beloved smartphone, a miracle device that lets you do almost anything as long as it involves poking clumsily at large icons. The public is writing rhapsodies about the latest video game, which looks absolutely amazing as virtual wallpaper but mostly consists of following simple instructions to earn stupid upgrades. The public is on Facebook or Snapchat or Twitter, where everyone and his second cousin just posted a link to a poorly worded argument that you absolutely have to respond to. The public is cheerfully feeding the gossip machine one kind of information, over and over and over–here’s the easiest, cheapest, quickest, simplest way to get our attention–and the engineers have no choice but to respond. Because they built this goddamn thing. And now they’re stuck inside it, with the rest of us.