The most recent attack on women and minorities in Silicon Valley has arisen unexpectedly from Google. Mounted by an anonymous Google engineer as a “manifesto”, it presents no facts, regurgitates disproven theories on the “biology” of men and women and, most tellingly, blames diversity for upper management’s cancellation of underperforming products at Google.
There are a lot of women who have worked on technology projects in SV over the years (me included), but you wouldn’t know it because no one writes about it, so no one believes that it happened even though this is a young industry and most of us are still alive. That missing piece of the story leads to the notion that women have not had any involvement in any technology and it’s a man’s world. It’s an absurd notion.
Whenever one sees these attitudes one also sees history has been deconstructed to focus only on one person at the expense of others – unless earlier in the history of the field there were key women who could not be deconstructed, like physics has Curie and Meitner. Those who control the information – tech journalists, writers and amateur enthusiasts – have had an almost laser-focus on men. Why?
Many people have already started to address the clear absurdities of this person’s claims, from both an internal and external perspective, along with myriad professional press musings too numerous to mention (try google).
Of special note, the key weakness in this memo is that women and minorities have nothing to do with the lament of Google “demoting the core business” and “killing beloved services”. This is odd, and leads to speculation that this “manifesto” is nothing more than a disguised attack on the streamlining decisions of Google CEO Sundar Pichai. In other words, this guy is trolling people with a red herring of “diversity”, and the real intent is to embarrass Google executives for cutting a “beloved” project. Move along folks – nothing to see here.
But the fact that many people see this rant as factual does merit some discussion. This means we have to dive into history a bit. I know most people have little patience for the past. But it does help to know how we got from here to there. It wasn’t random chance.
In practice, tech readers rarely notice the name of the author on the article, which is why it’s pretty easy to write about hard tech even if you’re a woman. But they do notice who is being interviewed, reviewed, or cited as an authority, and it’s usually a man because, as any editor or publisher will tell you, “That’s what the reader wants”. If this seems circular and under-justified, it is.
The second item is the current obsessive focus on low-level “pipeline” women in tech. While it’s important to get women into the system, it’s actually retaining them at the lead engineer, line manager and director level that matters. There is no focus on that.
Thirdly, it was not uncommon for women to be part of a Founding Team that was funded in the 1980’s. Startup teams were typically at least three people reflecting technology, operations and/or finance, and sales. Even if a woman was not an engineer, she would still be viewed as an equal Founding Member for her business, marketing or sales skills. This was also true of black men and women, especially in business and sales due to the their strong presence in old-line companies like IBM.
This Funding Trinity structure became less common as the Cult of the Geek became a meme in Hollywood as a follow-on to the Western antihero. The story was recast from a team of rather dull startup business equals making spreadsheets and chips and PCs to the lone tech-guy going against all odds to fight the System. It was amusing at the time, but it’s been done to death. How many men are going to write another Steve Jobs movie or opera? I’d rather see an opera about Marie and Pierre Curie. Now that was a scientific tour-de-force love story.
But it’s considered normal storytelling in journalism and entertainment to interview / romanticize / suck-up to men when anything serious is discussed to avoid alienating the reader aka men. Are only men readers, users and developers of tech? No, they aren’t. This unquestioned assumption perpetuates the notion that women don’t work in the field, aren’t interested in studying or reading about tech (or science, economics, politics …), and that men are the only instigators of creativity and change. It’s lazy writing, but it’s easier to meet Internet deadlines when you write by recipe rather than by the old-fashioned research / rewrite / review.
Coincident with this fascination of the lone geek, the tech people who rolled out of Berkeley and Stanford at that time found themselves in a rather unpleasant quid-pro-quo: to get a good reference, a student might have to spend considerable amounts of personal time on unpaid / low-paid tech projects. Since jobs were rather scarce (we went through several big recessions in the 1980s, kids), and a reference was really important to a decent job, there were lots of people willing to do this. It became a bit of a seller’s market. In a seller’s market, choices often become based on whim and comfort, and that’s exactly what happened. A like prefers like situation developed among key professors and their lowly student help to reduce management overhead and increase their collegial network. It’s human nature to seek familiarity and comfort when excellence is a commodity. This myth of “someone like me is easier to manage” prevails today.
There was also considerable selection bias in computer science and engineering majors in the 1980s. At Berkeley, there was so much demand for engineering one had to compete to enter as a freshman in the college, which precluded people who were unsure from entering the major. A woman who wanted to be an engineer had to not only know how to apply directly to the College of Engineering at Berkeley, but also have the confidence and will to be an engineer despite the high school tendency to channel women towards the “softer” majors (if encouraged to go to college at all).
The safety valve on learning programming was the few restrictions in the 1980s for non-major students taking courses in CS at Berkeley, and those could be easily waived by a Dean. Berkeley tightened this loophole in the 1990s due to budget cuts, essentially cutting out many people who the decade prior could still take CS courses while in other majors. This led to an EE/CS bottleneck. Stanford had a much smaller pipeline, as do most top-tier private schools. Berkeley was the big one for matriculating people in the field in SV, and its stranglehold on access had a profound impact for two decades.
This skinny-pipeline reduced-risk preference of “guys like me” also was the golden ticket to investment referrals. The hard tech innovation that flowed out of universities – from Berkeley Unix, to RISC, to databases, to languages – was a lucrative and exciting opportunity for people who resented the indentured servitude of academia. They left to fund startups based on these technologies. And the most skilled at these technologies were the same people who had been most willing and able to do work for nothing. Stanford, sensing an opportunity, actually refined the pipeline for investment, offering students access to alumni referrals and networks for a “piece of the action”, and reaped a windfall. Berkeley, in contrast, retreated further into academic narrowness, resenting the desertion of so many into the very industries they helped spawn.
The reason we are seeing discontent today has two key factors: 1) the ability to access excellent introductory and focused courses in programming at a cost-effective level is within the reach of many and 2) the value of a EE/CS degree has declined. The latter is a result of SV growing to encompass mature industries and verticals. Unlike twenty years ago, it isn’t particularly important for every programmer to know how to write a compiler or understand graph theory, and many excellent programmers are self-taught, strangely enough just like many of the early SV pioneers. Most programmers and engineers also work on extant projects, adding some code here, fixing a bug there, and rarely work on a new project or technology de novo. There is more demand for Stanford business school graduates to manage logistics and funding than Berkeley CS programmers to create new technologies.
In addition, the reliance on global access to talent has had an unexpected effect. The number of women from other countries with STEM degrees working in companies in SV is quite high. The majority of women I meet at women-in-tech events (and by this I mean hard tech since that is my field) is dominated by foreign researchers, programmers and engineers from India, China and former Eastern Bloc nations. There is far less stigma for a woman to go into a STEM field in these countries, and it shows in practice. These women are educated, ambitious and not afraid to speak out.
The American women I see at technology events are most commonly clustered in the data science area, and often possess advanced degrees in STEM fields. They are comfortable with data science because many STEM fields work with very large datasets and the tools, techniques and processes are the same when one is analyzing weather patterns or consumer patterns. There is also a reemergence of the value of biology, physics and mathematics degrees in biotech, aerospace, and fintech, respectively. In all cases, the calibre of talent is high and increasing.
The preference of companies like Google for obtuse whiteboard quizzes from upper-division CS classes over that of work, references and experience to validate “fit” reinforces a “CS-degree from top-10 university” bias that is obsolete in industry today. It also has the effect of favoring hiring of recent college graduates over those who have more experience.
Most of the tech pioneers – women and men – who actually did accomplish interesting projects / research / startups / technologies in the prior generation would be weeded out of the hiring pipeline today because while they had a heck of a lot of experience working on technology projects, they didn’t spend their time studying code quiz books. I have a Berkeley physics degree. While it’s a plus to people like Elon Musk (who also has a physics degree) in emerging industries of new space or electric vehicles, it is a minus at Google, FaceBook and other SV new old-guard companies. That is how their metrics and processes work, and they’re happy to keep it this way.
But are they really happy? Is this stasis good for their business? SV management has clearly not kept pace with the social changes in our industry, preferring nostalgia and a “that’s how it’s done” attitude to on-the-ground knowledge and change. This is the same pattern that emerged in the prior generation of old-guard companies of Xerox, Bell Labs and IBM, among others.
The hard truth is many successful SV companies are stuck in a midlife crisis where doing things the old way and fitting in is more important than challenging extant processes, technologies and business models. When this occurs, the time is ripe for a paradigm shift. This is now happening, and it’s making a lot of folks very uncomfortable. They lash out. They blame others. They want thing to go back to the way it was.
To sum up, the requirement of unpaid labor and selection bias by EE/CS professors on key projects at Berkeley and Stanford in the 1980s, the tightening of the pipeline due to budget cuts in the 1990s, reducing the ability of men and women not already declared in the major to “try” programming, the increased reliance of investment on innovative technology startups through this narrowed academic pipeline via referral, and the tech press fascination and support of male enthusiasts who reinforced a Cult of the Geek led us to what we see today – a peculiar devout belief that programming is a man’s job. And that belief is threatened by the sheer number of women in SV now clamoring for a seat at the table.
Sweeping aside all the vanity, programming at its core is working with words in a stylized manner to achieve a desired function. I’ve always found programming more akin to writing a sonnet in terms of the structure than prose. Fixing code is like writing a limerick. It’s not male. It’s not female. It’s just a tool, no more male or female than a pencil. We spend a great deal of time teaching kids in school to learn the tools of language, writing, mathematics and science. Programming is just another tool, with no special or endowed gendered significance.
I think I’ll go write a sonnet. It’s been a while, but I still know how.