Hacker News new | ask | show | jobs
by look_lookatme 2232 days ago
That's not what he said. He said on average women do not prefer the field, and optimizing to change that observable fact is a fool's errand.

I don't have an opinion on whether he's right or wrong but I don't think he shouldn't be able to make that argument...

3 comments

> He said on average women do not prefer the field, and optimizing to change that observable fact is a fool's errand.

Yes, it still leaves open whether the following is true:

> when women do prefer the field, they tend to be amazing engineers

Hence I'd say there is nothing wrong with the statement.

Not just women but anyone. For any social group,the share of that group may never be at or close to their make up of the population. People come from an insanely diverse set of upbringing and social expectations.

If they think 50% of engineers should be women, I would like to hold them to an even higher standard by requiring that out if that 50% of women,each national origin,ethnicity,political/ideological upbringing and social status (mother,single,divorced,etc...) is also proportionally represented. If this is about making up for historical prejudice then let every women of a wronged group be equally represented. You will soon see the silliness of the pursuit.

It is the cause that needs policing and adjustment (upstream) not the effect.

> Hiring practices which can effectively lower the bar for “diversity” candidates by decreasing the false negative rate

Damore deliberately chose to cast doubt on whether the women Google hired were as good as the men. Otherwise, it's easier and less controversial to just argue that any form of discrimination is wrong.

It doesnt cast doubt though..

You cant have so much more women than the rest of the industry without that being true. There literally is not enough women coming through engineering schools to allow for that without you a) discriminating against men based on sex, and b) lowering the bar for women.

There is no doubt. It is a mathematical given.

I see what you mean I think but it's really not a mathematical given. If you have two groups where group 1 with size S scores X with variance V==0.4X on average in a given skill (score means competence), and group 2 with size S/2 scores 0.8X with variance V on average; then it's possible to recruit a set of people from group 2 and the same number of people from group 1 so that the average competence level of people hired from group 2 is higher. All you need to do is to pick more carefully (invest more resources, etc.).

So even if we accept that the average female engineer is not as good as the average male, and there are less female engineers than male, and that Google hires more female engineers than other players in the industry (relative to male hires), it is still absolutely possible that the bar is not lower. Maybe they look harder, maybe they pay more to female engineers with a given competence level, maybe there's no difference, etc.

I dont think you understand the hiring pool available to mountain view at all.

The rate of female attendance in comp sci has been in decline since the 70s. At my university it was quite literally 100 to 1.

You cant have a 50-50 hiring pool without some fuckery there.

And for what its worth i absolutely do not accept that the avergage female engineer is not as good as the average male engineer.

If anything, they are anecdotally better equiped.

> All you need to do is to pick more carefully (invest more resources, etc.).

So what amounts to increase their total comp because theyre female and you have a quota to fill?

Thats sexual discrimination is it not?

I really need to say while im passionately pissed off about the way identity politics has infected this industry i am more than welcome to more females joining the industry and more than hapoy to see the broculture fuck right off.

What University has 100 to 1? That's way, way more unbalanced than the typical cs program.
> All you need to do is to pick more carefully (invest more resources, etc.).

This needs fleshing out. One goal is to give equal preference to equally deserving candidates, say candidates with individual score x, regardless of their group membership. It's not clear that you can pick based solely on score x and still achieve that goal, no matter how carefully you pick. If you're picking on x and group membership, you're harming one group of people with score x (assuming a fixed headcount, where advantaging one person means disadvantaging another).

It’s not hard to imagine that Google simply hired an enormous outsized chunk of the top female talent via outreach and just making vastly better offers to industry leading women.
At the time, a Googler explained here how they did it without doing either a or b https://news.ycombinator.com/item?id=14957764
Their description of what recruiters do seems like discrimination. They said that recruiters would deliberately go out of their way to find diversity candidates, instead of treating everyone equal. That matches up with a leaked email where the recruiters were told to stop processing candidates who weren't diversity hires.
Dropbox gives recruiters a diverse bonus equal to the difference between hiring an entry level IC1 and an IC6 (senior staff engineer) [0].

[0]: https://news.ycombinator.com/item?id=22741024

Google and where i work both benefit greatly from having increased diversity. It massively improves business outcomes over the monoculture culture we used to strive for.

My take is that when you optimize for equal outcomes instead of equal opportunity you break something at the opportunity end (its either or, not both.)

Regarding a or b, this is mostly b, lowering the bar. But its also a, sexual discrimination.

If you think its not, please explain how you think it would go down if google was focusing hiring efforts to hire white males because they wanted to match the current male to female ratio seen in engineering schools and currently have too many females in the office.

Because i fail to see how aiming to hire more females than the industry schools can bare simply because there are too many men in the office is anything but sexist.

Maybe one day people can stop looking at immutable physical characteristics like sex and skin color. Until then it's all discrimination by another name.
At Google casting doubt was a good in the early days. Even in search relevance of the results were the most important factor, even if it meant much less clicks, and the culture that Eric Schmidt created with the founders reflected that.

Nowadays I think it's a better strategy at Google to not discuss if you don't agree with its policies. It's still a great workplace as long as you can discard politics.

> It's still a great workplace as long as you can discard politics.

There are large swathes of people who can't just "discard politics" because they are directly affected by the current political status quo.

Sorry, but if you're talking about Googlers, that's bullshit, they're fine. 200k/yr, 300k/yr+ worth of fine.

It's great to have empathy for people in a less fortunate position, and yes we have a lot of problems in our society, but FAANG software engineers are arguably the least oppressed people on the planet.

Is this a class reductionist argument? Are you saying, for example, a black women at Google doesn't deal with adversity and is no longer effected by politics because they make a large amount of money?
Everyone is affected by "everything". Especially politics. People discussing it means they need to acknowledge how policies affect them and how it affects others, and is it good or bad "in general".

Or is that a too naive view of what Googlers are supposed to be capable of?

Right now a hard working black woman would have much harder time kicking herself out of Google than a white man. Also women are especially encouraged to go for promotion, and as a white man it just made me want to put in the work to go for promotion less there.

What I learned over time is that it's just better to work 8x5 days, don't do overtime/weekends, because it's not worth it anymore. For 8x5 hours of work Google is amazing.

> FAANG software engineers are arguably the least oppressed people on the planet.

Heterosexual cisgender white christian male FAANG software engineers may be. Racism and other bigotry doesn't stop affecting you because of economic class or other social position of advantage orthogonal to the basis of the bigotry.

Similarly, police officers are also a relatively privileged group of people, but 90% of mistaken identity shootings of off-duty police by police between 1982 and 2010 were shootings of black or hispanic officers.

> 90% of mistaken identity shootings of off-duty police by police between 1982 and 2010 were shootings of black or hispanic officers.

That's an interesting statistic. Can I have the source? (This is a genuine question, I'm not trying to cast doubt on what you said.)

That's actually not what "false negative rate" means - what he's trying to say is that you can find more candidates which are just as good, because you're putting more effort into searching for them. But of course it's an easy thing to miss, especially if you're not familiar with that terminology - and with existing tech-hiring practices that accept a higher false-negative rate to begin with. Damore never expected his memo to be spread as widely as it was, and this ongoing confusion is part of how it became such a political hot potato.
"cast doubt". That's a weasel word, because he said nothing of the sort.
And the lady engineers should put out some great work in reply and refute him.
That's not good.

Imagine how you'd feel. Imagine people would generally claim that someone who looks like you is a bad engineer (say, someone with your hair color). And then imagine that people would expect you to constantly disprove that assumption, just because of your hair color.

You'd be on the edge all the time. You'd never feel at home in the field. And whenever you interact with someone who doesn't have past work experience with you, they'd still assume you're a bad engineer.

I can't understand how doing great work isn't a great thing.

As the moderation bears witness, we live in an age where positive is somehow negative.

But don't you think that the fairer the selection (blind to things such as gender, race or politics) the more confidence you'd have that you deserve to be in the position you are? How does a woman engineer should feel, knowing she was given a boost in the selection process by some kind of affirmative action? Isn't that in itself a source of uncertainty?
Have a lot of female programmers told you about the unfair advantages they feel they have in the field?
Not sure what you mean. All the female programmers I've come across had been hired without receiving any boost from an affirmative action program. Had I worked in a company that gave a score boost to candidates belonging to any group X, I could have the legitimate doubt that without the boost some of the people in the group X wouldn't have been hired. It's simply a necessary consequence of the affirmative action.
Originally you were presenting an opposition which is based on a theoretical objection held _by women themselves_, one which you've now confirmed you've never actually seen. If you're going to advocate on people's behalf, limit it to concerns they've actually expressed themselves.

You've now rephrased your objection that you would have a problem because you would assume female programmers were less qualified, which is not what you said in your original comment.