Hacker News Reader

Comment by krainboltgreene | Hacker News Reader

> You are hired to build product, not obsess over the diversity of the employee base, unless you are hired in a DEI role specifically.

pssst those two things are connected

> Discrimination happens against an individual.

This is just weirdly wrong?

> If you see discrimination happen against an individual, and you privately and discreetly address it with the appropriate people (could be with those directly involved, with superiors, with HR, whatever seems appropriate in that situation), that is not a political act

This is an incredibly political act.

View on HN

It's hard not to escape the conclusion that the goal of Basecamp's new policy is to quietly suppress or ultimately evict people who think like you.

There is no connection whatsoever between diversity and building product. Note: you haven't tried to explain your claim there is, just implied it's obvious, probably because you can't. I've seen people try to honestly intellectually defend a connection between employee diversity and product design before, and it's always a joke. You get anecdotes about face recognition not working on black people (reality: the software in question just had trouble with low contrast images and also struggled with white people's faces). And that's about it. I can't recall ever seeing someone actually argue that more women == better product because, well, that's a pretty sexist thing to believe isn't it? It implies men have no empathy, or that there's a universe of product requirements women will only tell to other women, etc.

If discrimination is happening against an individual and you work with the management chain to resolve it, that's not a political act, because you aren't attempting to radically remake the entire organisation for everyone (politics), but only address the specific problem that was actually observed.

There is an absolute wealth of connection between more diversity and better products: it’s not that a monoculture can’t make products for everyone, it’s that it’s easier for a team that has wider representation to do so: it’s a competitive advantage for them.

Want examples beyond hand driers? Apple missed menstrual cycle tracking in the Health app. NASA offered to send 1,000 tampons with the first woman in space - and had to abandon its first all-woman spacewalk for bad planning in kit availability. Google Photos tagged black people as gorillas. (Regardless of the cause, it shipped). SnapChat shipped a yellowface filter. Google Translate delivers highly gendered results when translating from gendered languages into English. Google Home is 70% more likely to recognise male voices because of training set bias. Women are 49% more likely to be injured in car accidents because the seating position for crash testing is designed for men. And - yes! - Pinterest has struggled to reach men as users apparently because of its SEO strategy.

Want more examples? Ask any passing woman what app she uses that clearly didn’t think about women.

Diverse teams make better products. The stats are out there.

Trying to argue this highly political. We wouldn’t even be able to talk about your point at Basecamp. Does that not worry you? Don’t assume they will always agree with your politics just because they’ve done this.

Another (potentially fatal!) example: https://www.sciencedaily.com/releases/2020/08/200812161318.h...

> Women are more likely than men to suffer adverse side effects of medications because drug dosages have historically been based on clinical trials conducted on men, suggests new research from the University of California, Berkeley, and the University of Chicago.

It's a pretty subtle distinction but I'll try to make it.

The examples you give are mostly product examples (i.e. the product lacking features or not correctly addressing its target audience). Therefore I think it would be appropriate to discuss those in the workplace, but NOT as a function of e.g. the gender make-up or racial make-up of the team, but rather as a function of competence in addressing those issues.

In particular, imagine a hypothetical scenario where we are building a product, and we receive the feedback that many women find it difficult to use, and we want to fix that by hiring an expert.

The goal is to hire someone who understands marketing this type of product to women. Although it's maybe likely the most suitable candidate for this would be a woman, it could also possibly be someone else - and we would be looking for experience/evidence toward competence in that area, regardless of the candidate's gender or other characteristics.

To make the conversation apolitical, it would be about specific competencies and lacks of the team or product, NOT about the team itself having "not enough short blue-eyed Turkish women over 50" or some other group.

Also, to keep the conversation apolitical, the conversation would not extend to broader social questions outside of the specific product or problem. It would not stray into generalizations about groups. It would not appeal to emotional arguments like "safety."

None of my examples needed experts to address — just someone of the affected group on the team and (yes) comfortable enough to speak up. Also, even in your hypothetical, there will still be issues that are difficult to raise internally because the entire area is political. And so nobody will, until your company embarrasses itself in very public ways.

People bring politics and emotion. If your customers or staff are people, it is more effective to be able to discuss their issues directly without contorting yourself into a "just focus on the work" correctness.

Is the implication here that Apple had no women working on the Health app, Google Photos has no black team members (and if they did, they would have explicitly checked this case?), Snapchat has no asians, etc? Or that some magic ratio would have caught these issues?

There's some serious assumptions buried there that don't pass the smell test.

Yes, that is absolutely the implication. These mistakes are so egregious they're almost instantly obvious to people who know to look. For instance, the Health app tracked seemingly everything under the sun, including blood-alcohol levels. The reaction from women when it shipped was instant: where's the period tracker?

If there was a woman on that team it says even less good things about Apple that they shipped without it.

But: Apple's diversity report from then [1] says they were 80% male in tech, so while the assumption might be serious, I don't think it's unreasonable.

[1] https://www.theverge.com/2014/8/12/5949453/no-surprise-apple...

Apple case is somewhat odd as it's a guilt by omission. I would bet large amounts of money on there being at least woman within the product chain on this (director, PM, TPM, designers). That is, I don't think there's as strong of a correlation between presence of an identity and attention to features/products they would be affected by. For example, I could see a PM reading articles like this [1] and deciding there's too much potential heat on a period app in the default offerings. FWIW I also think this should have been there, and Apple can position themselves nicely w.r.t privacy concerns.

Like, at its core, I agree with the statement that having women on the team would not decrease the probability that these sorts of mistakes don't happen. But its positive impact has to be non-deterministic (or else getting into some implications about <identity> as a group which I can't get behind).

[1] https://www.vox.com/the-goods/2018/11/13/18079458/menstrual-...

If there was a woman on that team it says even less good things about Apple that they shipped without it.

So you admit it's nothing to do with diversity, then, just an ambient assumption that the needs of women should automatically trump the needs of men and women combined?

The simplest explanation for Apple's decision is not that they're all raging sexists who need to be forced to hire more women by activists, but rather that when adding features to an app they are ranked by how useful they'd be to the entire userbase. Feature development is a zero sum game - adding that means not adding a different feature. Any feature exclusively for women will have half utility by that obvious metric and lose automatically to features useful for everyone. So the real question is why would anyone expect a general purpose health app to have a period tracker? Isn't that a pretty grotesque sense of entitlement by a small handful of women?

As for the others, come on. That's exactly what I mean by ridiculous arguments. Any photo app should have a black team member who is tasked to spend all day taking selfies and not, say, fixing bugs or adding features? If some company was dumb enough to actually do that, they'd just get criticised in other ways, because such black-testing jobs would be menial low wage jobs, not highly paid and respected software engineering positions. Some of those examples aren't even oversights, like Apple's straightforward prioritization decisions or the fact that Google Translate uses gendered language when translating languages that use genders pervasively. They're just stuff a small minority of activists have chosen to get faux-offended over. And the few that remain are clearly not caused by lack of diversity awareness given you chose to cite Google Home (I'll take your word for it on the stat), a product from one of the companies most visibly committed to filling the ranks with women. They even hired a black female AI ethics researcher, so obviously it's not lack of diversity that caused the training set problems.

In other words none of the examples you cite have anything to do with "diversity" and cannot, because as you admit, you have no idea what the demographics of those teams is. You're just assuming that they must be all made up of un-caring men who never think about others, which is exactly the sexist assumption underlying all identity politics that I highlighted!

Periods are relevant to a "small handful" of women? Have you discussed your thinking with any women?

Is it a "pretty grotesque" sense of entitlement at work when a man looks for a urinal in a public building? Or when a disabled person looks for the wheelchair ramp? These things have half utility at best by your "obvious metric".

Take this further: imagine architecture firms were staffed mostly (80%+) by women. They keep getting complaints because their doorways are only 5'11" high (ideal for 99% of women's bodies), they don't provide urinals and so on.

Is it a wild assumption that having more men – or, god, "experts in catering to men (who might be men but also women)" - on their teams would help address these repeated own goals? Or do we need to spend years getting the data that shows us the real, mysterious, non-political underlying cause of this gender bias in buildings?

It seems to me that buildings and medicine and cars and gym equipment and everything, need to be usability tested on all groups of people: short, tall, men, women, black, white, thin, fat, old, young, wheelchair, blind, deaf.

But I don't think all those groups of people need to take part in the construction work.

It'd be nice with more diversity in tech. So please don't misunderstand. I just think that there are other better examples of how diversity is good.

(Usability testing: Testing early prototypes I suppose -- it's not that easy to redo, say, a car, once it's in production.)

> keep getting complaints because their doorways are only 5'11" high (ideal for 99% of women's bodies),

That thought makes me feel upset! (Although I'm not a human. But I am tall) Thanks for a good example

I feel like you're deliberately mis-interpreting me. It's a small number of women who kick up a big fuss on Twitter about the choices of Apple's PM team.

Reply on news.ycombinator.com