Casting out the heretic at Google

My Facebook friends are talking about the Google programmer’s memo regarding why there aren’t more coders at Google who identify as “women” and “Non-discriminatory ways to reduce the gender gap.” Here’s a representative comment:

female VC: Even if it were true that there were “population level” differences in women/men, Google doesn’t hire nearly enough people to make this relevant. Back when I worked at Google, it was a haven of rationality. Said man should be fired for inane use of statistics.

her friend: Did you read the entire document?

female VC: No, just the first two paragraphs. It was very poorly written.

The article quotes the Google VP of Diversity:

Part of building an open, inclusive environment means fostering a culture in which those with alternative views, including different political views, feel safe sharing their opinions. But that discourse needs to work alongside the principles of equal employment found in our Code of Conduct, policies, and anti-discrimination laws.

The guy’s boss:

… we cannot allow stereotyping and harmful assumptions to play any part. One of the aspects of the post that troubled me deeply was the bias inherent in suggesting that most women, or men, feel or act a certain way. That is stereotyping, and it is harmful.

The Code of Conduct is available online. Will this guy ultimately be fired for violating “Each Googler is expected to do his or her utmost to create a workplace culture that is free of harassment, intimidation, bias, and unlawful discrimination”? Among other things the guy says women are prone to “Neuroticism (higher anxiety, lower stress tolerance).This may contribute to the higher levels of anxiety women report on Googlegeist and to the lower number of women in high stress jobs.” [Update: these are precisely the parts of the memo and the Code of Conduct cited by CEO Sundar Pichai in a company-wide response email.]

Separately, is the memo ridiculous on its face? If an employer is short of Worker Category X, isn’t the first explanation that these workers aren’t being offered sufficient pay? Google could have plenty of female workers in any category if they were willing to pay up, no? Why would an intelligent hard-working woman want to be a coder at Google, get paid 1/20th the cost of a decent house nearby, and stare at a screen all day when she could instead be a physician, get paid 2X the cost of a decent house near her office, and interact with people all day? (Alternatively, she could realize the spending power of a programmer by having sex with a couple of programmers.)

Programming is considered by most people to be a disagreeable boring job, a desk-bound analog of sewer cleaning or garbage pick-up, so, absent much higher pay to women, wouldn’t we expect there to be a similar male-heavy gender ID ratio as in other disagreeable jobs?

I asked the Facebookers in the thread why Google didn’t stop with the fine sentiments about diversity and instead just pay whatever it took to get the workers it wanted. One guy said “It would be illegal to adopt practices that specifically aim to increase pay for women.” (But why not instead pay for some characteristic that women tend to have in a larger quantity than men?) Another responded with “A very small data point but I sat on a small committee which placed a few hundred elite tech grads at Google and FB. In this limited sample size of high performers, women were making 40% more.” (i.e., the market is working; women programmers are more valuable to employers and therefore can command higher pay)

Readers: What do you think? Is this guy a pinhead? Will Google fire him? Once his name gets out, will any other employer be willing to take the risk of hiring him? (thus opening themselves to a slam-dunk sex discrimination case because they have created a “hostile environment” for female coders)

[Update: I broached this topic with a female programmer friend. She said that she thought there was some truth to the heretic’s point of view, e.g., that men were more willing to put in hours of dreary solitude when learning to code. I was able to convince her, however, that the only real problem was money. I ran through a list of mutual acquaintances, all of whom were smart, capable, and had great jobs. She agreed that all of these women would have been able to become competent software engineers and would indeed have done so if the compensation and working conditions were competitive with the career options that they’d actually chosen (e.g., medical specialist, Wall Streeter).]

[Update 2: A friend sent me this post from Slate Star Codex: “About 20% of high school students taking AP Computer Science are women. … which exactly matches the ratio of each gender that eventually get tech company jobs.”]

[Update 3: “No, the Google manifesto isn’t sexist or anti-diversity. It’s science,” by Debra Soh, a professor in Canada with a PhD in sexual neuroscience, says “the memo was fair and factually accurate. Scientific studies have confirmed sex differences in the brain that lead to differences in our interests and behaviour.” Thus the Hillary supporters at Google who mocked Trump voters for being “science-deniers” now find themselves denying neuroscience whose implications they don’t like. (Separately, I must part company with Professor Soh on a non-scientific point. She says “seeking to fulfill a 50-per-cent quota of women in STEM is unrealistic.” Give me a stack of cash and I will fill any quota!)]

[Update 4: In response to a Facebook posting characterizing the memo as containing “logical errors,” her friend responded with “I’d love your thoughts on what you think is illogical. You may not agree, but to me it was all a reasonable position to state, and I agree with most of it. I’m a woman in tech, and a woman CEO, and I know full well why there aren’t more women in my position and it has less to do with discrimination or bias than biology and lifestyle preferences. Would love your thoughts as to which bits you didn’t agree with?” (note that the response kind of proves the Google Heretic’s point; she softens her disagreement with “would love your thoughts”).]

[Moderator is removing some comments to keep the total within the 50-comment display limit of Harvard’s software.]


the reason I left is that I came into work one Monday morning and joined the guys at our work table, and one of them said “What did you do this weekend?”

I was in the throes of a brief, doomed romance. I had attended a concert that Saturday night. I answered the question with an account of both. The guys stared blankly. Then silence. Then one of them said: “I built a fiber-channel network in my basement,” and our co-workers fell all over themselves asking him to describe every step in loving detail.

At that moment I realized that fundamentally, these are not my people. I liked the work. But I was never going to like it enough to blow a weekend doing more of it for free. Which meant that I was never going to be as good at that job as the guys around me.

35 thoughts on “Casting out the heretic at Google

  1. In bidding for new CS grads from elite universities who’ve gone through our internship program (40% female) it’s quite consistent that the women get better offers from Google, Uber, etc. Both in cash and intangibles: allowing them to work from their preferred office (NYC is popular), or putting them into the exact group they’re interested in.

    We extend offers to nearly all former interns as the program is quite selective, but end up with only 25-30% new female hires per cohort due to being outbid more regularly on women.

    That said, the manifesto was weak and unnecessarily insulting. La Griffe du Lion put out a slightly more convincing statistical analysis years ago:

  2. If google is in a liberal bubble, is there downside to them being there?

    I think there’s a perception that tech companies only political risk, comes from the left, from being sued by SJWs and the like

    I’m not especially sure that’s true, I don’t have any particular opinion of whether using a heavier regulatory hand towards facebook or google would be helpful or wise (or even if that’s an accurate report of Bannon’s intentions)

    I do think though, the more stories like this come out, and the more the google doodle makes people eyes roll, the more people feel no special affinity for these companies, and the more political cover anyone wanting to regulate them, (or make them pay any other sort of political price), will have

  3. Phil:
    Why is it that nearly everything you write must now contain reference to paternity shenanigans vis-a-vis women who could get rich off the system by sleeping with the right man? I get that this is a hugely important and personal topic of yours, and one where I vehemently agree that things are really quite fucked up.

    The problem is that mentioning it in nearly every piece you write – often with the most tenuous of tangential connection – often undercuts an otherwise cogent take on the core topic. It is to a point where if you were to publish a Twinkie review, you would put an aside to say “(If a woman wanted to be able to afford a nearly unlimited supply of these delectable confections, all she needs to do is meet a man in Manhattan, conceive of a child with him in Wyoming, and sue for custody from an address Antarctica.)”

    Phil, I love you’re take on things, but this topic of yours is really subtracting from otherwise astute, contrarian work.

  4. Greg: Thanks for the suggestion and I’m sure you’re right in general. For this particular post, though, which is about female labor participation, family law is relevant.

    The late 20th century change in family law to no-fault divorce and guaranteed 50/50 property division caused millions of women to withdraw from the workforce or reduce their hours. See research from University of Chicago by Alessandra Voena, referenced in

    I’ve seen a lot of articles talking about how the percentage of female programmers pre-1990 was higher than today. 1990 was when child support guidelines put in place by states (directed by federal law), thus making out-of-wedlock child-bearing profitable (see ).

    Americans who can collect child support work fewer hours per week than similarly situated Americans who can’t. In California, 94 percent of the people collecting child support are women (see ). Being a coder at Google requires longer hours than the average child support collector wants to work. cites data showing that single mothers who could collect welfare based on having a child cut their working hour compared to married women ineligible for welfare. Why wouldn’t we expect women to make the most dramatic cuts in jobs that were never enjoyable to begin with?

  5. Considering we all work for Google, in a sense. Are we all going to require a VP of diversity & to fill out diversity license agreements to go along with the thousands of pages of license agreements required to compile an Android program? Wouldn’t what’s being sold as artificial intelligence do a better job as a VP of diversity than it’s doing as a personal assistant?

  6. Irony, of course, is that there is a market force pushing software companies into leftist positions (since influential demographics in the industry are strongly pro-left; also this creates positive feedback loop). In this context Google’s behavior is rational, and this guy tried to fight it with reason and argument (I was laughing hard when I was reading the manifesto) and got steamrolled in the process, proving the argument made in his memo by personal example.

    I work at Google and I can tell you that effort spent on brainwashing and fighting of thought crimes is huge. All republicans and libertarians I know keep quiet about it at work.

  7. SK: Since you work at Google and could have seen the original… did the original posting by this guy include hyperlinks to journal papers? The way that it has been printed by Gizmodo makes it look as if he just wrote down his personal opinions, but it reads a lot more like a summary of published research. It would be unusual for someone with a PhD in science to write something without references (but if one were to cut and paste from a browser window into an email to a journalist, maybe references that were hyperlinks would disappear).

  8. While there may be a lack of local female talent (20% you say?), perhaps Google can simply hire Indian and Chinese women in sufficient numbers to achieve the desired parity.

    Let me as an aside register my strong displeasure at a Harvard alumnus being fired.

  9. This whole thing about “diversity”, “inclusion” and what have you, is being overblown.

    Would the “diversity” VP at Google want the best programmer coding the controllers of the Radiation Therapy machine that she might need to use to treat her or family or would she settle down for “diverse” programmers to work on the code?

    Would the “diversity” VP at Google take a business flight on an airliner with poor maintainers record simply because she is required to be “diverse” in her travels?

    What about companies and colleges encouraging employee and students to stand-out and speak-out their mind? Is this true only when one does so in private?

    Those are not a new issues, but the internet and fast-news and impatient society that doesn’t know how to slow down isn’t helping.

  10. Firing the guy for voicing an opinion — civilly! — whatever it may have been, just proves that “diversity” is code for “anti-conservative sentiment.” I work in a place where tremendous amounts of time and energy is taken to pontificate on the same message. We’re a social justice company that just happens to make a physical product. I fear enough for the same reaction to my own opinions that I’m posting this anonymously.

  11. Would the “diversity” VP at Google want the best programmer coding the controllers of the Radiation Therapy machine that she might need to use to treat her or family or would she settle down for “diverse” programmers to work on the code?

    The good news is that Google doesn’t deliver direct radiation at tumors. It delivers advertising to eyeballs.

    Firing the guy for voicing an opinion — civilly! — whatever it may have been, just proves that “diversity” is code for “anti-conservative sentiment.”

    You could just say he was expressing stereotypes civilly and perhaps eloquently. A line has to be drawn somewhere.

  12. My, how the definition of “firing offense” has drifted! OTOH, why would a tech professional put this in writing instead of just dripping it out at the juice bar?

  13. How is this markedly different than the firing of Larry Summers (about which Philip blogged at length) for his women & science speech? Around the same time, Summers had the audacity to state that developed countries’ unwanted garbage should be sent to Third World countries who are interested in generating revenue (in theory to cover necessities for their improverished populations — whereas in practice this likely lined the Swiss bank accounts of their dictators). The 20% gender ratio is actually more generous than I’ve personally observed at 1 Hacker Way in Menlo Park among actual coders/programmers, as there are non-engineering jobs on that campus which skew more heavily female than the overall Silicon Valley workforce. Wondering whether tech companies try to use the employees in the cafeterias & laundry facilities to boost their official female employment nos?

  14. In the original memo on page 2, the author says
    “Of course, I may be biased and only see evidence that supports my viewpoint.”
    The author does acknowledge that he maybe biased. He does have references, but it would be good to have more references to actual population studies, most of the references are just to other media articles.
    Why does the media or some academic provide a good counter argument?
    Is it because he is guilty of that hideous crime of making a political incorrect argument?

    Why would women want to go into a field to compete with a bunch of underpaid and overworked male code monkeys with no social skills that can be replaced with more cheaper code monkeys? Women are much smarter than men, that is why they go to study medical fields and law.

    We should be more worried about equal pay for doctors and lawyers.


  15. There’s apparently a huge bias against women in masonry(0.1% females), drywall installation(0.3%), diesel engine repair(0.5%). That’s where the low hanging fruit of anti-female bias is!

  16. Why does google allow non-work related discussions at workplace at all? I believe if it does and even encourages them it can be sued for firing for such posting, or for not providing in-house censorship services if such are demanded by google policy. Is it to keep employees married to their workplaces and forget that when longer someone works quality of code diminishes? I think everything google does can be accomplished in 8 hour day productively and rant or trolled after work on own time, especially by people who can write demanding algorithms in short time period allotted for interview, unless it secretly building quantum cloud from scratch.

  17. Perhaps Google could achieve it diversity goals by implementing an in-house employee selective breeding program just like the San Diego Zoo for endangered species. Google could establish a pay-to-play breeding program until the desired result is achieved over a few generations.

  18. Phil:
    As others said, there were links, it was originally published as Google Doc, as most of documents at Google are. I guess PDF is “export to PDF” version of this doc.

  19. @Suzanne

    I asked a lady who is an electrician and owns an electrical contracting firm. Her primary culprits in order are:
    1) The unions
    2) The trade schools
    3) The generally boorish behavior of electricians at work

    The union slots are increasingly inherited. The fathers get slots for the sons. Probably the biggest sex leveler is that fathers are starting to get slots for daughters who want to be electricians. Eventually there will be mothers getting slots for daughters and sons too.

    Trade schools who dare not antagonize the unions, and share the preferences for sons. The graduates with only the public school training have a very hard time getting apprentice positions that lead to the necessary licenses. Even though many take a non-union job, the trade schools are influenced by the unions.

    On the job there is a lot of rude, crude commentary of all sorts. You either accept it or find another job. It’s OK to object when it’s aimed at you or someone you know. It’s not OK to object in general.

  20. Instead of firing him, he should have been promoted to Senior Firmware Upgrade Developer in the CoSentry – Sioux Falls Data Center in South Dakota.

    Ideally he would be reporting to a woman there. And a genderfluid individual named Pat.

  21. @rjh — your account is consistent with what an air traffic controller (female) fellow lapswimmer told me back in the 1990s. The male controllers were so crass and crude (not unlike the Mooch, I suppose) that she sued the FAA, in part due to deliberate flatulence by her all-male colleagues. Her suit wasn’t entirely successfully (granted, it’s difficult to win employment suits against the federal govt) as she was demanding to cut back to part-time in order to spend more time with her 3 boys. (See Ziskie v. FAA.) You’re right that going forward there should be more females in the trades, and in fact, the president’s latest push toward that goal should abet the efforts of parents to push their daughters in the direction of skilled “blue collar” professions. However, in my view, Summers (whose background is as a research economist) was looking at raw data, where he found that at the top tiers of STEM fields, where things get more complex, women weren’t necessarily able to excel in the same numbers as men.

  22. > You could just say he was expressing stereotypes civilly and perhaps eloquently. A line has to be drawn somewhere.

    So that line is simply “expressing a stereotype,” regardless of civility or eloquence?! Regardless of the stated desire for an objective, reasonable outcome, and in the absence of any clear malice? Again, you’ve proven the author of the memo’s point: we can’t even have a reasonable discussion about the issues any more. One isn’t even allowed to question the liberal position, even in the nicest way possible. It simply isn’t permitted. Diversity of everything but thought, apparently.

    4 Scientists Respond, saying the author got the facts right: (Link is dead. Mirror:
    The memo didn’t address this paradox directly, but I think it’s implicit in the author’s critique of Google’s diversity programs. This dogma relies on two core assumptions:

    * The human sexes and races have exactly the same minds, with precisely identical distributions of traits, aptitudes, interests, and motivations; therefore, any inequalities of outcome in hiring and promotion must be due to systemic sexism and racism;
    * The human sexes and races have such radically different minds, backgrounds, perspectives, and insights, that companies must increase their demographic diversity in order to be competitive; any lack of demographic diversity must be due to short-sighted management that favors groupthink.

    The obvious problem is that these two core assumptions are diametrically opposed.

    But that’s really all beside the point, isn’t it? People just bring their biases and prejudices to the issue, and whatever it is they WANT to be true, facts and statistics and 10,000 years of human history be damned.

  23. There is clearly a genetic basis for intelligence. A head of lettuce is way less intelligent than the average human (although many posters on this blog may prove the exception to this empirical observation).

    However any scientific definition and understanding of ‘intelligence’ beyond the most basic generalities is so weak that the spectacle of people bashing each other about how to link it to chromosomes would be funny if it didn’t also come along in tow with such a large suitcase of sexism and racism.

    Until we have an understanding of intelligence which is detailed enough to demonstrate implementation of human level performance in a computational system, rather than the laughably and hopelessly vague and medieval definitions used by psychologists, these debates are pointless. If you can’t even define what you’re measuring, then how about getting back to us when you can.

  24. It occurs to me that Danielle Brown, VP of Diversity at Google, would have been PERFECTLY positioned to have a discussion on how the memo got it wrong, and persuade people that the position the memo argued against was actually the more enlightened path. But she didn’t do that. She demured, and then they fired him.

  25. today’s “Opinion” section of the WSJ in a piece titled, “Memo to a Google Engineer” echoes what Larry Summers said over a decade ago (which got him sacked), viz., “The distribution of traits within a population says nothing about the traits of any particular person.” When I asked my son about gender diversity within his engineering team in Silicon Valley (which at the lunch hour, was mostly nerdy Asian-American males from a cursory look, and the 2 bosses I’ve met have been white males not quite old enough to be my twenty-something son’s father), he did note that ” . . . they’re trying to add more women, and they did recently add a woman.” His gf was there for this discussion, and she has enough background in calculus & economics that she is using large data sets to route out Medicare fraud for a consulting firm, but she doesn’t code like Karlie Kloss. I joked with my husband afterward that it’s great for her to know her bf is surrounded by other men.

  26. so Summers was forced to resign (wasn’t outrightly sacked) according to Wikipedia, and Sheryl Sandberg (who had worked for Summers at Treasury, at Google at the time of this statement) came to his defense: In January 2005, at a Conference on Diversifying the Science & Engineering Workforce sponsored by the National Bureau of Economic Research, Summers sparked controversy with his discussion of why women may have been underrepresented “in tenured positions in science and engineering at top universities and research institutions”.

    Summers had prefaced his talk, saying he was adopting an “entirely positive, rather than normative approach” and that his remarks were intended to be an “attempt at provocation.”[39]

    Summers then began by identifying three hypotheses for the higher proportion of men in high-end science and engineering positions:

    The high-powered job hypothesis
    Different availability of aptitude at the high end
    Different socialization and patterns of discrimination in a search[39]
    The second hypothesis, the generally greater variability among men (compared to women) in tests of cognitive abilities,[40][41][42] leading to proportionally more males than females at both the lower and upper tails of the test score distributions, caused the most controversy. In his discussion of this hypothesis, Summers said that “even small differences in the standard deviation [between genders] will translate into very large differences in the available pool substantially out [from the mean]”.[39] Summers referenced research that implied differences between the standard deviations of males and females in the top 5% of twelfth graders under various tests. He then went on to argue that, if this research were to be accepted, then “whatever the set of attributes… that are precisely defined to correlate with being an aeronautical engineer at MIT or being a chemist at Berkeley… are probably different in their standard deviations as well”.[39]

    Summers then concluded his discussion of the three hypotheses by saying:

    So my best guess, to provoke you, of what’s behind all of this is that the largest phenomenon, by far, is the general clash between people’s legitimate family desires and employers’ current desire for high power and high intensity, that in the special case of science and engineering, there are issues of intrinsic aptitude, and particularly of the variability of aptitude, and that those considerations are reinforced by what are in fact lesser factors involving socialization and continuing discrimination. I would like nothing better than to be proved wrong, because I would like nothing better than for these problems to be addressable simply by everybody understanding what they are, and working very hard to address them.[39]

    Summers then went on to discuss approaches to remedying the shortage of women in high-end science and engineering positions.

    This lunch-time talk drew accusations of sexism and careless scholarship, and an intense negative response followed, both nationally and at Harvard.[43] Summers apologized repeatedly.[44] Nevertheless, the controversy is speculated to have contributed to his resigning his position as president of Harvard University the following year, as well as costing Summers the job of Treasury Secretary in Obama’s administration.[45]

    Summers’s protégée Sheryl Sandberg has defended him saying that “Larry has been a true advocate for women throughout his career” at the World Bank and Treasury. Sandberg described of the lunch talk “What few seem to note is that it is remarkable that he was giving the speech in the first place – that he cared enough about women’s careers and their trajectory in the fields of math and science to proactively analyze the issues and talk about what was going wrong”.[46]

  27. ““Larry has been a true advocate for women throughout his career”

    And yet he was excommunicated nevertheless.

    A similar story happened to an acquaintance of mine who was a political science professor at one the Boston area schools. In his entire career, he was as progressive, “in thoughts, words, and deeds”, as one could possibly get. Despite being a tenured employee at his university, he was accused of creating a hostile environment and forced out. That was in 1995 or thereabout.

    In comparison, the current case is not terribly interesting and will soon be forgotten even sooner than Summers’s has already been. Just another victim of the progressive SJWs for whom Marx/Lenin/Stalin/Mao/Brezhnev are their natural spiritual leaders.

  28. My criticism of Summers was the same as of the Google Heretic: why isn’t it as simple as “women can get better jobs at higher pay”? I pointed out that working as a physician for 10X the salary of a science post-doc was not an obvious sign of reduced IQ (see )

    It seems to have taken him a whole year to resign:

    At that point I wrote up and said “Summers was deservedly castigated, but not for the right reasons. He claimed to be giving a comprehensive list of reasons why there weren’t more women reaching the top jobs in the sciences. Yet Summers, an economist, left one out: Adjusted for IQ, quantitative skills, and working hours, jobs in science are the lowest paid in the United States. This article explores this fourth possible explanation for the dearth of women in science: They found better jobs.”

    Programming at Google isn’t quite as horrible a career as academic scientist (free lunch!), but I think the same questions should be asked.

  29. “fourth possible explanation for the dearth of women in science: They found better jobs”

    Well, without hard numbers, e.g. the science candidate pool size, acceptance rate, objective acceptance criteria (probably non-measurable in principle, not in the last place due to ideological climate du jour unless one employs the Chinese Imperial Examination system), the fourth explanation is a mere conjecture. Just as Summers’ list of reasons was.

    More importantly, I think that the real or perceived inadequacy of gender related occupational representation research by these two persons is not the most important lesson one can learn from their respective plights and other plights of similar nature.

  30. Adjusted for IQ, quantitative skills, and working hours, jobs in science are the lowest paid in the United States. This article explores this fourth possible explanation for the dearth of women in science: They found better jobs.”

    So that must be the difference between the genders. Men don’t respond to economic incentives.

Comments are closed.