DEBRA SOH — No, the Google manifesto isn’t sexist or anti-diversity. It’s science

No, the Google manifesto isn’t sexist or anti-diversity. It’s science

By now, most of us have heard about Google’s so-called “anti-diversity” manifesto and how James Damore, the engineer who wrote it, has been fired from his job.

Titled Google’s Ideological Echo Chamber, Mr. Damore called out the current PC culture, saying the gender gap in Google’s diversity was not due to discrimination, but inherent differences in what men and women find interesting. Danielle Brown, Google’s newly appointed vice-president for diversity, integrity and governance, accused the memo of advancing “incorrect assumptions about gender,” and Mr. Damore confirmed last night he was fired for “perpetuating gender stereotypes.”

Despite how it’s been portrayed, 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.

As mentioned in the memo, gendered interests are predicted by exposure to prenatal testosterone – higher levels are associated with a preference for mechanically interesting things and occupations in adulthood. Lower levels are associated with a preference for people-oriented activities and occupations. This is why STEM (science, technology, engineering and mathematics) fields tend to be dominated by men.

We see evidence for this in girls with a genetic condition called congenital adrenal hyperplasia, who are exposed to unusually high levels of testosterone in the womb. When they are born, these girls prefer male-typical, wheeled toys, such as trucks, even if their parents offer more positive feedback when they play with female-typical toys, such as dolls. Similarly, men who are interested in female-typical activities were likely exposed to lower levels of testosterone.

As well, new research from the field of genetics shows that testosterone alters the programming of neural stem cells, leading to sex differences in the brain even before it’s finished developing in utero. This further suggests that our interests are influenced strongly by biology, as opposed to being learned or socially constructed.

Many people, including a former Google employee, have attempted to refute the memo’s points, alleging that they contradict the latest research.

I’d love to know what “research done […] for decades” he’s referring to, because thousands of studies would suggest otherwise. A single study, published in 2015, did claim that male and female brains existed along a “mosaic” and that it isn’t possible to differentiate them by sex, but this has been refuted by four – yes, four – academic studies since.

This includes a study that analyzed the exact same brain data from the original study and found that the sex of a given brain could be correctly identified with 69-per-cent to 77-per-cent accuracy.

Of course, differences exist at the individual level, and this doesn’t mean environment plays no role in shaping us. But to claim that there are no differences between the sexes when looking at group averages, or that culture has greater influence than biology, simply isn’t true.

In fact, research has shown that cultures with greater gender equity have larger sex differences when it comes to job preferences, because in these societies, people are free to choose their occupations based on what they enjoy.

As the memo suggests, seeking to fulfill a 50-per-cent quota of women in STEM is unrealistic. As gender equity continues to improve in developing societies, we should expect to see this gender gap widen.

This trend continues into the area of personality, as well. Contrary to what detractors would have you believe, women are, on average, higher in neuroticism and agreeableness, and lower in stress tolerance.

Some intentionally deny the science because they are afraid it will be used to justify keeping women out of STEM. But sexism isn’t the result of knowing facts; it’s the result of what people choose to do with them.

This is exactly what the mob of outrage should be mobilizing for, instead of denying biological reality and being content to spend a weekend doxxing a man so that he would lose his job. At this point, as foreshadowed in Mr. Damore’s manifesto, we should be more concerned about viewpoint diversity than diversity revolving around gender.

Debra Soh writes about the science of human sexuality and holds a PhD in sexual neuroscience from York University.

Sundar Pichai Should Resign as Google’s C.E.O.

Sundar Pichai Should Resign as Google’s C.E.O.

There are many actors in the whole Google/diversity drama, but I’d say the one who’s behaved the worst is the C.E.O., Sundar Pichai.

The first actor is James Damore, who wrote the memo. In it, he was trying to explain why 80 percent of Google’s tech employees are male. He agreed that there are large cultural biases but also pointed to a genetic component. Then he described some of the ways the distribution of qualities differs across male and female populations.

Damore was tapping into the long and contentious debate about genes and behavior. On one side are those who believe that humans come out as blank slates and are formed by social structures. On the other are the evolutionary psychologists who argue that genes interact with environment and play a large role in shaping who we are. In general the evolutionary psychologists have been winning this debate.

When it comes to the genetic differences between male and female brains, I’d say the mainstream view is that male and female abilities are the same across the vast majority of domains — I.Q., the ability to do math, etc. But there are some ways that male and female brains are, on average, different. There seems to be more connectivity between the hemispheres, on average, in female brains. Prenatal exposure to different levels of androgen does seem to produce different effects throughout the life span.

In his memo, Damore cites a series of studies, making the case, for example, that men tend to be more interested in things and women more interested in people. (Interest is not the same as ability.) Several scientists in the field have backed up his summary of the data. “Despite how it’s been portrayed, the memo was fair and factually accurate,” Debra Soh wrote in The Globe and Mail in Toronto.

Continue reading the main story

David Brooks
Politics, culture and the social sciences.
Getting Trump Out of My Brain
Can People Change After Middle Age?
Before Manliness Lost Its Virtue
Jeff Flake Plants a Flag
JUL 28
How Cool Works in America Today
JUL 25
See More »


John Conroy 16 hours ago
The confirmation bias by Brooks here echoes Damore’s. Irony-impaired Damore posts his crackpot retrograde treatise that women are emotional…
Katie 17 hours ago
I don’t know David. I read his memo, and while I agree the tone was amiable, he lost me when he claimed that women don’t handle stress as…
Nat irvin 17 hours ago
Gosh, easy to say when all you have to do is write a column weekly AND when you’re fed up with Trump decide to ignore him. Pichai has a very…
Geoffrey Miller, a prominent evolutionary psychologist, wrote in Quillette, “For what it’s worth, I think that almost all of the Google memo’s empirical claims are scientifically accurate.”

Damore was especially careful to say this research applies only to populations, not individuals: “Many of these differences are small and there’s significant overlap between men and women, so you can’t say anything about an individual given these population-level distributions.”

Newsletter Sign UpContinue reading the main story
Sign Up for the Opinion Today Newsletter
Every weekday, get thought-provoking commentary from Op-Ed columnists, the Times editorial board and contributing writers from around the world.

Enter your email address
Sign Up

You agree to receive occasional updates and special offers for The New York Times’s products and services.

That’s the crucial point. But of course we don’t live as populations; we live our individual lives.

We should all have a lot of sympathy for the second group of actors in this drama, the women in tech who felt the memo made their lives harder. Picture yourself in a hostile male-dominated environment, getting interrupted at meetings, being ignored, having your abilities doubted, and along comes some guy arguing that women are on average less status hungry and more vulnerable to stress. Of course you’d object.

What we have is a legitimate tension. Damore is describing a truth on one level; his sensible critics are describing a different truth, one that exists on another level. He is championing scientific research; they are championing gender equality. It takes a little subtlety to harmonize these strands, but it’s doable.

Of course subtlety is in hibernation in modern America. The third player in the drama is Google’s diversity officer, Danielle Brown. She didn’t wrestle with any of the evidence behind Damore’s memo. She just wrote his views “advanced incorrect assumptions about gender.” This is ideology obliterating reason.

The fourth actor is the media. The coverage of the memo has been atrocious.

As Conor Friedersdorf wrote in The Atlantic, “I cannot remember the last time so many outlets and observers mischaracterized so many aspects of a text everyone possessed.” Various reporters and critics apparently decided that Damore opposes all things Enlightened People believe and therefore they don’t have to afford him the basic standards of intellectual fairness.

The mob that hounded Damore was like the mobs we’ve seen on a lot of college campuses. We all have our theories about why these moral crazes are suddenly so common. I’d say that radical uncertainty about morality, meaning and life in general is producing intense anxiety. Some people embrace moral absolutism in a desperate effort to find solid ground. They feel a rare and comforting sense of moral certainty when they are purging an evil person who has violated one of their sacred taboos.

Which brings us to Pichai, the supposed grown-up in the room. He could have wrestled with the tension between population-level research and individual experience. He could have stood up for the free flow of information. Instead he joined the mob. He fired Damore and wrote, “To suggest a group of our colleagues have traits that make them less biologically suited to that work is offensive and not O.K.”

That is a blatantly dishonest characterization of the memo. Damore wrote nothing like that about his Google colleagues. Either Pichai is unprepared to understand the research (unlikely), is not capable of handling complex data flows (a bad trait in a C.E.O.) or was simply too afraid to stand up to a mob.

Regardless which weakness applies, this episode suggests he should seek a nonleadership position. We are at a moment when mobs on the left and the right ignore evidence and destroy scapegoats. That’s when we need good leaders most.

Moving from Nikola to Hugo

Moving from Nikola to Hugo

As part of my recent Move to Hugo I wrote a few small tools that may be useful (with tweaking) for someone else doing the same, or moving from a similar static hosting platform.

Basic configuration

I was extremely happy to find that I could keep my existing permalink structure just by editing config.toml:

post = “/:year/:month/:title/”
Importing content

The first part of content was easy, as Nikola and Hugo have very similar methods for storing static files. I just had to copy the contents of my serialized-nikola/files tree to my serialized-hugo/static tree.

The next part was a little trickier. Nikola:

Supports Restructured Text, which I used for a few posts
Has a different frontmatter format (Restructured Text style, wrapped in HTML comments for markdown)
Has a different date syntax – actually, supports more flexible dates, so I had several syntaxes in play.
I had enough posts (137) that doing this by hand would have been not fun at all, so of course, I scripted it. It’s not a generically useful script – I even hard coded in some of my paths – but if you’re doing a similar migration it might be a good starting point. Nikola has a crazy-high degree of flexibility, this script specifically only handles the subset of what I was using.

Please also note, for these one-and-done scripts I tend to ignore my typically rigorous testing and error checking habits. 😁

Here’s the process:

conversion process

Yet again, pandoc to the rescue, as it made converting from Restructured Text to Markdown a breeze.

    args = ['pandoc', '--from=rst', '--to=markdown', '--output=-']
    data['content'] = subprocess.check_output(args)

I went through the majority of the posts by hand, and there were only a few things that got left behind (that I noticed), like YouTube embed codes, that were easy to fix up by hand. It was really incredible to run the script and in a matter of seconds have the livereload refresh to reveal a fully functional site.

Spring Cleaning

While I was migrating, I realized there were a lot of images and random other files which were no longer used, many from posts which I had retired long ago. Almost all static site generators (including hugo) do struggle with image/post locality; there’s a good discussion in a github issue.. Because of this, I had about 500 files in my static/ directory, and I had no idea which were still being referenced or not.

Thanks to all the posts being in markdown, I realized the paths would have to show up in those files, so built out a simple tool to

find all the files in static/, and normalize the paths to match what they actually look like from the webserver
Open every file in the content tree, and keep track of any of the static file paths which appear in them
Output all the static files which have zero references
This tool is in go: unused_images.go.

Go’s, unsuprisingly, very powerful when doing this kind of task. A snippet showing the gathering process:

var seen map[string]int

func findImages(path string, f os.FileInfo, err error) error {
imageRe := regexp.MustCompile("images/.*$")
seen[imageRe.FindString(path)] = 0
return nil

func main() {

seen = make(map[string]int)
filepath.Walk(images, findImages)


It ran seemingly instantly, and spit out a list of over 100 files that could be deleted, which is excellent (and 100 images fewer to have to check state on every time the deploy/sync process runs.)

How fast is your site? You can test here theperformance of any of your sites from across the globe

How fast is your site? You can test here the
performance of any of your sites from across the globe

This test measures how long it takes to connect to your site and for one page to fully load. A very important value to pay attention is the “time to first byte”, which gives us how long it took for the content to be sent back to browser to start processing the page. If you are using a CDN (Content Delivery Network), your connection time may be low, but if the time to first byte and total time is also not low, it will not give you much in terms of performance.

Sucuri® is a registered trademark of Sucuri Inc. in the United States and/or other countries.