Can you explain why hiring women who are just as good as, if not better than, men at the job, but who would prefer (on average) to work elsewhere, perhaps somewhere with more women employees, would make the business do worse?
And can we note that I replied to someone saying "he didn't say that", suggested a place where he did and got replies from different people saying "yes he did say that, and he's correct" and replies saying "no, you're misreading what he said as it's explained in more detail later".
So it appears at the very least (and I feel I'm being generous here), that his words can be misread by people on both sides of the issue, which might be worth bearing in mind for those trying to paint his detractors as a hysterical mob who haven't even read the memo.
Can you explain why hiring women who are just as good as, if not better than, men at the job, but who would prefer (on average) to work elsewhere, perhaps somewhere with more women employees, would make the business do worse?
In the section you quoted, he doesn't say that hiring more women is bad for business, he said that doing so via discriminatory hiring practices, and if you hadn't ellipsized that line, there would have been an explanation for why right there. The full bullet point reads "Discrimination to reach equal representation is unfair, divisive, and bad for business." So he says that unequal representation is partially explained by gender differences in interests, and then that trying to eliminate the differences, specifically via the method of discriminatory hiring, is bad for business because it's unfair and divisive.
To be clear, the reason I'm interpreting the "differences" he refers to as differences in interest as opposed to competence is because of the content in the "Personality differences" section of the memo, which is the only place he addresses specific gender differences. In that section he says that women are less interested on average in tasks which focus mainly on systemizing, and are less interested in high-stress jobs, as well as that they are less likely to advocate for their own promotions.
I replied to someone saying "he didn't say that", suggested a place where he did
No, you replied to someone saying "he didn't say that" and suggested a place where he said something ambiguous — the tl;dr, which is necessarily ambiguous.
So it appears at the very least (and I feel I'm being generous here), that his words can be misread by people on both sides of the issue
Who do you think is misinterpreting what? I've pointed out what I think you're misinterpreting and offered an alternative interpretation. You don't get to simply assert that other people are misinterpreting the memo, you've got to offer some evidence of that.
He didn't say "unfair and divisive, and therefore bad for business", he said "unfair, divisive and bad for business". My ellipsis doesn't change the meaning, unless we are, once again, falling back on him presenting his ideas in an ambiguous manner. Which, even if true, makes misinterpretations his responsibility.
In the memo he literally asks for changes to make Google more welcoming for conservatives. He even claims this will be good for business. How do you square this with your contention that any kind of discrimination, even for positive purposes, will lead to negative outcomes due to it being unfair and divisive.
Does it make sense for him to be claiming both these things? I'm not sure that he is even saying that (though it seems tone deaf at the least to present both ideas in the same document).
Also, I'm really not sure that saying women score higher on neuroticism and therefore can't handle stressful jobs as well, or are introverts and so have trouble leading is purely about interest. I'm not sure where this defence about "only talking about interests" comes from, I don't get that from the document. It's certainly not explicitly spelled out.
I'm coming to the conclusion that people aren't just misinterpreting but have actually invented a totally fictitious memo that they find easier to defend than the real one.
You cannot forcibly hire someone who would prefer to work elsewhere (if that someone can really work elsewhere, e.g. not be worried with matters of survival and quality of life). This is probably a reason that in a lot of developing countries (e.g. India) gender gap in IT is much smaller than in developed ones (e.g. Norway, #1 gender equal country)- salary is choosen over comfort.
Now, imagine a situation: 100 persons (80 men, 20 women - ratio observed by actual HR) are looking for a job in Google, which has 10 open positions for them. Google wants to hire the best possible candidate - and hires 8 men and 2 women; both genders are choosed from top 10% of candidates of a same gender (Assuming their skill does not correlate with their gender). That hiring policy does not reduces observed gender gap between candidates at all. So let's introduce positive discrimination (so called affirmative action), and say, that Google wants to hire 5 men and 5 women (1:1 ratio, but even 7:3 will suffice to show worsening effect, albeit in weaker proportions). Now men are choosed from top 6.25%, and women - just from top 25%. The is no gender gap anymore, but now there is a considerable skill gap - top 6.25% men are having considerable advantage over 25% women. More to that: men are not happy because they think they have struggled much more for the same position than women; women are not happy because they think they were hired not for their skills, but for their gender. Also,both men and women being paid based on their performance alone will result if considerable wage gap.
The original memo stated the same, albeit in less transparent form: personal skills cannot be measured with 100% accuracy, so there is always a chance to hire underperforming person (false-positive) or not to hire well performing person (false-negative). Positive discrimination radically reduces false-negatives, which inevitabely magnifies number of false-positives (see Type I and Type II errors in statistics and Neyman-Pearson criterion). Author proposed not to fight on-enter gender gap with discrimination, which would result in negative effect for company's performance, but to make company's environment to be a more attractive (which may reduce on-enter gender gap, reduce number or women leaving industry permanently and won't compromise performance)
>Can you explain why hiring women who are just as good as, if not better than, men at the job, but who would prefer (on average) to work elsewhere, perhaps somewhere with more women employees, would make the business do worse?
The issue is in the act of finding them. This is where statistics and discrimination comes into play. Lets say that ratio of women vs men that are interested in working at google (and capable of being hired) is 20% to 80%. Then if your goal is to reach gender parity, you must do one of three things: (1) test significantly more women vs men to find those 20% at a higher rate (which involves active discrimination against men), (2) lower the standards specifically for women so that the pass rate for women is higher, or (3) active intervention along gender lines to increase the percentage of women that are interested and capable of passing a google screening. It should be obvious that 1&2 leads to worse outcomes for Google. (3) doesn't directly lead to worse outcomes, but it could be seen as divisive to discriminate against men in these outreach programs, which could indirectly lead to worse outcomes for Google. Now, one could argue that these gender-based outreach programs are just to counter cultural sexism that have kept women from being more interested in engineering jobs at Google. But if the stats in the memo are accurate, at least some portion of the disparity is not due to external factors but individual choices based on inherent (dis)interest. If this is true then such outreach will largely be unsuccessful for the cost in money and resentment.
And can we note that I replied to someone saying "he didn't say that", suggested a place where he did and got replies from different people saying "yes he did say that, and he's correct" and replies saying "no, you're misreading what he said as it's explained in more detail later".
So it appears at the very least (and I feel I'm being generous here), that his words can be misread by people on both sides of the issue, which might be worth bearing in mind for those trying to paint his detractors as a hysterical mob who haven't even read the memo.