Very true.

It is a vague and meaningless concept. It has never been defined except by referring to antiquated notions of “race” stemming primarily from 18th and 19th century ideas.

In modern times, we’ve apparently added biological sex, sexual preference and a vague (and meaningless) concept of self-identity that can, apparently, cross all other definitions with impunity - even changing fluidly throughout the day.

To appoint people as “head” of this mess of ideas point to deeper issues in the organization.

The fact that categories are often fuzzy does not make them useless as a general heuristic for finding people who are more likely to share similar experiences with each other.

Almost every kind of categorization system shares the problems you talk about: people can call themselves "programmers" for a variety of different reasons, even though they're working in wildly different environments and might be totally unsuited for each other's jobs. People self-identify with political parties, even though there are no strict rules about what a "Democrat" or "Republican" are, to the point where beliefs of members of those parties might overlap with each other or even be the opposite of what the party normally espouses. We separate games/movies into genres, even though genres are often a complete mess that don't always map well to player preferences or experiences. We take a light spectrum and we break it up into distinct colors, even though the number of colors and lines between those colors are completely arbitrary and meaningless.

It's good to recognize that categories are fuzzy human inventions that are prone to abuse, and it's good to be cautious about putting a lot of weight on them in every situation. It's good not to assume that everyone in a category is identical. But abandoning categories entirely would make it much harder (even impossible) to go through normal day-to-day life.

So it is similarly useful to be able to think about categories like "Black", "Gay", "Conservative", "Christian", "Rural", "Trans" -- even as we acknowledge that those categories are just human inventions and that they blur around the edges. Those categories are still be a useful heuristic for trying to find people who are more likely to share certain attributes or who have had certain life experiences.

> Those categories are still be a useful heuristic for trying to find people who are more likely to share certain attributes or who have had certain life experiences.

In other words, prejudice is a "useful heuristic".

> In other words, prejudice is a "useful heuristic".

If you don't see any difference at all between community and othering, or if you think we need a single universal metric that defines whether categories are being used for evil in every single scenario, then I don't think you understood what I was saying about fuzziness.

Categories are a neutral tool, it's how they're used and the effects of their usage that determines whether or not an action involving them is problematic -- just like every single other tool in existence.

You are trying to rationalize the usage of “black” and “white” while simultaneously seeming to agree with my point.

I would say this to you: if you believe a diverse set of skills, experience and personality are important (I believe they can be) then use categories for that.

Skills categories are judged from education and practical work tests.

Experience is judged from CV and references from earlier employments.

Personality can be assessed from tests and judged during interviews.

These aren’t perfect either, but they are far more relevant than: “oh, you seem brownish, how nice”

Multiple issues:

- First, this is not an either/or situation. You can assess personality, skills, and experience in traditional ways, while simultaneously understanding that having a diverse collection of genders, races, ages, disabilities, etc on a team is often a competitive strength. You don't have to throw away people's CVs to do any of this.

There's an assumption that I see sometimes that caring about diversity during hiring necessarily means ignoring everything else, and that's just not what people like me are saying to do at all.

- Second, the role of a head of diversity should go way beyond simple hiring decisions, it can also mean trying to make a more welcoming environment that accommodates diverse workforces after they're hired, it can mean trying to explain why certain demographic outcomes and trends exist in hiring, and trying to figure out on a high level if there are areas where Google is biasing its hiring process or culture to be less diverse.

It should be kind of obvious, but Kamau wasn't just sorting applications based on skin color -- you wouldn't need multiple people and a "head" to do that. A diversity initiative is supposed to look at the entire company, identify subtle problems in multiple areas/departments, and come up with creative solutions to those problems.

- Third, I think you're discounting life experience in general as a contributing factor towards how people make decisions. It's not "you're brownish, that's nice." It's "Black people want to use our products, is there someone in the room who's going to notice if our AI mislabels them in photos?" It's "we want our UI to be usable by blind people, it would be nice if we had anyone in our office that can identify certain pitfalls." It's "we're designing a social network feature that will accidentally make doxing much easier, what are the odds someone who's LGBT+ in our office will raise their hand and point out how dangerous this is for their community?"

Those aren't theoretical concerns, they're issues that companies like Google care a lot about for very practical reasons. So it's a big jump to act like political diversity, age, life experiences don't impact how products get made. Understanding diversity means both acknowledging that these communities are fuzzy and that not everyone in them is identical, and also acknowledging that on net people in these communities have experiences and perspectives that can be valuable to incorporate into decision making processes.

- Fourth, you're missing that many diversity initiatives in companies don't exist in a vacuum. Google has a diversity initiative because it has problems with diversity, because there are unreasonable biases in its hiring processes and recruitment strategies against older, POC, and other minority applicants. Diversity initiatives in companies like this don't just randomly appear out of nowhere. They're not trying to get rid of good hiring practices, they're a response to existing flaws in the hiring process.

It's all well and good to argue that hiring processes should be egalitarian and that diversity will naturally arise if we ignore the categories. But the current outcomes aren't egalitarian, and ignoring the categories didn't lead to natural diversity. Argue if you want that the kind of diversity I'm talking about is irrelevant to quality, but doing so really only reinforces that the lack of diversity across the tech industry is abnormal and it's worth having teams inside of companies who try to figure out why their hiring outcomes seem to be biased in certain directions.

Why not? I can think of plenty of reasons why google would want a head of diversity.

- They may find diverse teams to perform better

- or maybe highly qualified people prefer to work on diverse teams so it helps with recruitment

- or their leadership value creating a more equitable society by giving people in under-represented groups the opportunity to work at google.

If any of the above are true why should they not have someone in charge of diversifying their workforce?

"Diverse teams" does not mean having a proper ratio of white, Hispanic, male, female, straight, gay or ginger. It means having people who have complimentary skills for a task and everyone got to that point differently. A legal team made up of one person from Harvard, one from UCLA, one from Princeton and one from a no name school is a type of "diverse". It's even more if one person was an engineer first, one was JAG, another was an accountant or some other profession before getting into law. A legal team where everyone is from Harvard and that's all they know since they've done nothing else with their lives, but check off different minority boxes, and have a token redheaded stepchild ginger, is not diverse. It's pandering due to both being lazy and being useless.

Someone that diversifies teams actually has to know the team well, what they do, and identity their strengths and weaknesses. Then, they need to figure out what kind of person can bring the neccessary perspective or skillset to round out the team... but that takes work. It's easier to just say, "You dont have a ginger on your team, here you go. Diversity achieved!" This is why people hate "diversity heads". They're glorified checklists.

> "Diverse teams" does not mean having a proper ratio of white, Hispanic, male, female, straight, gay or ginger."

Oh that's certainly what some people mean.. Some people have a blind spot when you bring up the lack of older people or from poor backgrounds.

They can be. Credentialism facilitates structural racism and class boundaries. Nobody gives a shit about gingers, that is a red herring.

One of my parents used to work in the housing discrimination space. Those same types of arguments were used there. Instead of credentials, they’d advertise in targeted ways, like only advertising at law or medical schools, etc.

Recruiters use colleges as a legal discrimination technique to target specific class origins. Unless you win the college lottery and get into Harvard, it is not possible for a kid from an urban setting to get the gig at McKinsey that is the entry screener used to get certain gigs. It’s like the old British civil service.

Companies are diverse as a side effect of hiring for competence. They aren't competent as a side effect of hiring with diversity as a goal. The statistically illiterate HR minions who preach the correlation between diversity and performance don't get this, because they aren't even educated enough to understand correlation/causation fallacy.

Totally disagree.

Even ignoring that diversity can be a goal in and of itself, and ignoring that a lack of diversity can be a useful metric for identifying other problems, there's also plenty of reason to believe that targeting diversity directly can yield performance improvements. If you're building a product for women, you should have some women on your team. If you're building a product for the general population, you'll get better results if your team has a variety of life experiences and viewpoints.

It's the same exact reason why it's good for a CEO to have experience in the industry that they're targeting with their products, and why managers who know at least a little bit about code works tend to do a better job of managing programmers. Perspective and life experiences absolutely help with decision-making processes and can help you avoid certain pitfalls in your product.

Diversity increases your available breadth of input into decisions and planning; it's like the human version of increasing code-coverage in your QA department.

Yeah, we're talking past each other about two different things now. Classic.

How did you interpret JPKab's comment? Maybe I misunderstood, do you agree that a focus on increasing diversity can lead to better products and better company decisions?

So are you claiming opportunity is equally distributed? I feel if you "hire for competence", you are going to end up with a highly concentrated group of wealthy people who went to the best schools.

Also, what is your metric for "competence"? You can train people to be computer programmers, but there are many other traits that are impossible to train. Like seeing the world from a perspective different than a middle aged white man, for one.

I think they are saying talent is equally distributed across all humans.

We know opportunity is not.

We know education is not.

Also, competence and "best" schools are not the same.

>Companies are diverse as a side effect of hiring for competence.

I mean, unless factors you evaluate for competence are just 'similarity to current decision makers'. Which is what people largely do in practice. This is why we see that tests show that hiring decisions vary dramatically on the basis of names, pictures, educational background, etc.

No one has a magic 8-ball that spits out an employee's power level. So something else is being measured.

But even if we COULD measure for competence - say in an objective numerical manner - the amount of times you'd optimize towards a local maxima rather than overall maxima would be significant.

So the measurement is flawed, and the algorithm is wrong. You'd think people on HN of all people would recognize the above as self-evident given how frequently 'I'm upset at tech hiring' threads are posted here.

When you’re making products for a diverse group of people, perhaps having a diverse team is required for competence.

This is probably the biggest reason, they are building products used by virtually everyone on the planet, they cannot afford mistakes that could arise just because their employees only represent a small percentage of human diversity. If all diverse groups were equally well-off, or all they understood the existence and power of privilege, or there was no risk of the built products putting certain groups at a disadvantage, maybe we wouldn't need to do this. But we know that's not true.

That's clearly not the motivation behind it otherwise age and political diversity would be similarly prioritized.

Age is somewhat prioritized in practice. Why do you think political diversity would be similarly prioritized?

There's a moral argument that comes into play with politics that doesn't for age or race. There's nothing even abstractly morally problematic about making your product appeal to women or old people or whomever. There may be reasons to avoid marketing or aligning your product to, say, people whose political views you disagree with.

We're not talking about products aimed at a particular demographic. We're talking about products with customers from diverse demographic backgrounds, where the argument is put that we should then have a diverse team so as to ensure we are serving that product competently to the diverse customer base.

My point is that this can't be the real reason for D and I initiatives, since if it was, there would be a similarly strong push to have equal political and age representation in order to reflect the reality of these products' customer demographics.

I don't think anyone is claiming that their priorities couldn't be improved.

The priorities won't change because the priorities were never stemming from an earnest attempt to represent customer demographics to begin with.

Not only is there no attempt to represent customer demographics on dimensions such as age, but there continues to be rank discrimination against people based on these dimensions, and the people pretending to be in favor of diversity are consistently and conspicuously silent on the matter.

And you're not educated enough to understand that "meritocracy" is a myth that entrenches existing power and privilege structures, because the people who are good at something are the people with the resources and social capital it takes to get good at that thing.

Skills and experience were not distributed fairly, but they are still real and still necessary. The fix for missing education is more education, not trying to somehow do without it.

Ahhh, so if I disagree with you on a point, it's due to lack of education?

That's a hell of a worldview.

I grew up in an impoverished family, in a trailer park, in a mostly black county. I know far better than you do about "privilege structures", but since I happened to be white, none of the diversity initiatives care or help.

Poor white people are underrepresented in tech. But the white elites think poor whites are in trailer parks because they are stupid. It's their fault, right? That's what all of you say, I've heard it said everywhere. "Trailer trash" this and that, said casually without batting an eye. Never questioning if maybe, just maybe, the kids born there have opportunities equivalent to yours.

It's obvious nobody cares. If they did, they would talk about it, or try to do something. But there are no activist professors for them, so crickets.

"Hiring for competence" doesn't automatically get you a diverse labor force, especially if "competence" is measured by metrics that are themselves demographically correlated to begin with (educational background from schools that aren't very diverse, experience at other firms that aren't diverse, etc.). Criteria like that can entrench established demographics, especially if the people who do the hiring look for similar backgrounds and traits that they themselves have... a bunch of WASPs hiring other WASPs because they come from the same schools or like the same governance tools or some framework, for example, or a bunch of Japanese businessmen excluding the white applicant who speaks perfect Japanese because they don't think he'll understand cultural norms, or a BLM group not wanting to hire the white diversity trainer because she ostensibly has worked mostly with Latino populations. There are always unspoken preferences, often cultural things, that bias hiring -- even if you aim for competence -- unless we blind the hiring committee and resumes, etc. A few studies about identical resumes with different last names (suggesting different races) suggest that competence is not really the best indicator of hireability.

> The statistically illiterate HR minions who preach the correlation between diversity and performance don't get this, because they aren't even educated enough to understand correlation/causation fallacy.

That seems kinda like a strawman? Who's arguing that? It's not like having a more diverse team will necessarily result in higher quality code, or more lines of code per month, or some similarly laser-focused metric of "performance".

Diversity can instigate cultural change at a company (do we value hard deadlines more, or should we emphasize work-life balance more?), make it easier to acquire and communicate with customers from diverse backgrounds (cross-cultural communications is hard for anyone, especially so for non-diverse teams with no specific training in it), fill some legal or marketing mandate (we gotta look diverse even if we aren't), help spur innovation in process or product development (what's "out of the box thinking" for one demographic may be a common process for another demographic), etc.

Of course there are costs too. Diverse teams may or may not work together as well, holidays might be different, food and beverage preferences may not line up with current offerings, religious or cultural conflicts may occur, staff polarization becomes more likely (along political/ideological/national origin lines, whatever)...

That's all to say diversity isn't just skin color but cultural backgrounds too, and you get all the pros and cons of that... for better or worse. It's far more nuanced than simply trying to hire for some arbitrary measure of "competence".

Why, because bias and prejudice wouldn't exist in such a world?

Well that would be too much to ask I think, those are natural mechanisms and typically not harmful.

I think positions like these are needed because there are severe forms of discrimination, hate, exclusion based on ethnicity, gender and so on.

Arguably, they're more about the "typically not harmful" things than about the severe cases.

The severe cases are bad, but they're also obvious. Your ordinary management should be able to handle that. They often don't, and it helps to have a special level of appeal when the chain of management fails, but that's not the real reason for the job.

The real reason is that those "typically not harmful" cases are cumulatively harmful. They're bricks in the briefcase of every employee being discriminated against. They get all of the usual problems of life, plus a new set aimed at them. So they don't perform quite as well, and aren't the best when promotion time comes around. Then you end up with a whole chain of command who thinks that those "typically not harmful" cases aren't the reason everybody in authority looks like them -- and then do nothing about it.

Dealing with the explicit cases is easier, even though companies often fail at that, too. If they can't handle the easy cases, there's nobody looking out for the hard ones. And worse, people often say, "Look, we fired the blatantly racist guy, why are you still complaining that every single manager is white? It's just a coincidence, OK?"

I was thinking in general terms of bias and prejudice. You can’t remove it from existence because those aren’t qualitative terms. A decent person can recognize and actively combat bias, but not erase it. You’re still going to have some form of bias. At least I haven’t met a single persom who does not.

Now you’re talking about discrimination. Those can appear mild and harmless in single instances, but the accumulation is the problem as you said but also what we could call passive ignorance. In my experience people are often not even aware of being discriminatory, some are even well meaning, but are patronizing.

But there is a key difference in a spectrum of bias/prejudice and flat out not reflecting on our behavior.

I realize I‘m discussing semantics here, just wanted to explain how I understand those terms in my response.