Hacker News Reader

Comment by AnaniasAnanas | Hacker News Reader

> since we need female candidates

May I ask why? Why does it matter if the candidate is male or female? Why not just hire the most fit candidates regardless of their gender and be done with it?

View on HN

   Why not just hire the most fit candidates regardless of their gender and be done with it?

Short answer, because you aren't actually capable of identifying them.

Longer: mostly because there is real value in team diversity, and it's almost impossible to remove your decision biases without actively addressing them. So if you identify a potential bias, it's not crazy to try and do something about it and see how that works for you.

There is a very real and pragmatic application for this very thing, specifically in AI work as identified by Fei-Fei Li:

https://www.wired.com/story/fei-fei-li-artificial-intelligen...

edit: my phrasing wasn't very clear. I mean for having a team from a diverse makeup.

Because that's how you get Apple Health omitting a monthly bodily function that a majority of the population has. Diversity of perspective makes a better product.

Is that really the reason why Apple Health didn't include menstrual cycle tracking? Or did they just decide not to add it in additional to a bunch of other things?

Not super familiar with that example but another one is all kinds of technology not working well with darker skin.

Early versions of facial recognition had difficulty working with darker skin due to genuine technical challenges, namely lower contrast between facial features and the background skin color (e.g. dark eyebrows on dark skin vs. dark eyebrows on light skin). This was not solved by hiring dark skinned engineers, it wad solved by a different technical solution: infrared cameras that could better sense features regardless of skin color for example.

The developers of facial recognition tools were aware of the early devices' limitations, and dedicating resources towards an approach that worked better on subsequent iterations. Why people think things would have automagically worked on the first iteration if the teams building the devices were more diverse, and why people continue to use this as an example of the necessity of diverse teams is a mystery to me.

It's in vogue in Silicon Valley to bias hiring standards in order to get more "equality" on the other side.

This has little to do with SV.

I've worked in SV and elsewhere and the push to hire disproportionate numbers of women in tech roles in definitely stronger in SV.

For example, at my company 27% of our tech workforce is women, as compared to an industry average of 20% (figures are from our head of diversity). Despite already having an overrepresentation, we continue to enagage in discriminatory hiring policies that favor women and we've bumped our target for women in tech roles up to 33%.

I haven't worked outside of SV in 5+ years but this level of aggressiveness in pursuing diversity is not something I witnessed elsewhere.

I work well outside of SV and the push for greater diversity is a large endeavour in most industries (I've worked in many).

I really disagree with your wording, and the slant you're taking with calling an above average representation in your company as an overrepresentation. You're comparing your number to the current standard which is often considered insufficient.

But I think we'll probably have to agree to disagree on that point, because this isn't the subject of the thread.

> I really disagree with your wording, and the slant you're taking with calling an above average representation in your company as an overrepresentation.

An above average representation is the definition of an overrepresentation. I'm not sure why you are attempting to state otherwise. Whether you desire a larger representation of women in tech as a whole does not alter the objective fact that a company with a share of women in tech roles larger than the tech industry average has an overrepresentation of women in tech roles.

The statement that the current representation of women in [this wasn't exclusively about tech, but you continue to represent it as such] tech is somehow a natural constant, rather than having been possibly influenced by any number of factors. I mean, you can't objectively use the current representation of women in [tech/other] as a baseline for what you're arguing. It has bias inherent—one that you're using to argue that going above said representation is somehow discriminatory. That's problematic—if objectivity is the ideal here.

Whether or not the percentage of women in tech is influenced by other factors was never stated in my comment, why you seem to think I am denying this is the case is a mystery to me. Regardless of said factors, this does not alter the fact that having a portion of tech workers greater than tech as a whole constitutes an overrepresentation. This is the literal definition of an overrepresentation.[0] To claim otherwise is subscribing to an alternative definition of an overrepresentation, and you can see why doing so will generate friction with other commenters.

In case this was missed in my original comment, my company does have explicitly discriminatory hiring policies that favor women over men (as well as favoring certain races). To call my company's hiring policies discriminatory is indeed an objective statement - regardless of whether you wish to call my company's representation of women an overrepresentation.

To clarify, my original point is that while many companies apply discrimination to balance out underrepresentations, it is quite aggressive to continue to apply discriminatory hiring policies even when the favored demographic is already overrepresented by a significant margin. I have rarely seen companies outside of SV go to the these lengths.

0. Merriam Webster defines overrepresentation as, "having representatives in a proportion higher than the average", https://www.merriam-webster.com/dictionary/overrepresented

I think you missed my point.

I was stating that your figure on the current representation of women in the industry may not be an objective representation but may be the result of many factors, and is not useful as a baseline for beginning to even determine what is over- or under-representative. Therefore, it can't be objectively considered.

But, that's all for me on this subject for now. All the best...

You still seem to be subcribing to an alternative definition of an overrepresentation. Whether or not factors influence the baseline population does not lessen the objectivity of measuring representation. You seem to be under the impression that the definition of an overrepresentation is, "having representatives in a proportion higher than some perceived ideal average." While under this alternative definition it may be fair to say whether not not a representation of women higher than the current average is a subjective statement, I am speaking in terms of the actual definition of the word which refers to representation with respect to the current average - not some perceived ideal average. When the proportion of women in tech is 20% and a company's proportion is greater than that, then an overrepresentation exists. The fact that you may think the proportion of women in tech may be greater absent certain factors does not affect the objectivity of that statement.

Reply on news.ycombinator.com