I'm an engineer at google and this bit about hiring someone because of their genes is false.
All candidates must past the same hiring bar, no one gets extra points for "diversity". So no, I never wonder why a coworker was hired.
All candidates must past the same hiring bar, no one gets extra points for "diversity". So no, I never wonder why a coworker was hired.
Retries, easier judges, removing some earlier barriers, etc would all be obvious discrimination in any other process.
You yourself admit that the process has false negatives. If those are in any way probabilistic (which is obvious that people get in after multiple tries), allowing minorities to try more frequently is lowering the bar for minorities.
If you let someone throw extra darts at the dartboard to hit the bullseye, that doesn’t make them as good as the people that hit it in fewer attempts.
> So no, I never wonder why a coworker was hired.
Good for you, but not everyone was so willfully ignorant when I worked there.
Yes, the company tells you that's how it works. But I was also a Google engineer, I did hundreds of interviews during a time when it was way less into diversity than it is today. And absolutely Google bent its policies to try and hire more women. Constantly.
Some tricks I saw used:
1. Assigning the best/most reliable interviewers to women. Recruiters confirmed to me that they did this.
2. Allowing women who failed the phone screens to proceed to on-sites anyway.
3. Organising recruiting events which banned men, e.g. there was since early years Google CodeJam, a coding competition that was pretty openly used as a way to find candidates. But the men crushed the women, repeatedly. I think there was a female finalist once in 13 years. So they set up a second CodeJam with large cash prizes, but banned men from it.
In recent times there was the recruiter who quit and claims he was told to not hire non-diverse candidates. I believe that. I was told about recruitment's diversity goals by recruiters themselves.
But still, there was a lot of resistance to bending the hiring pipeline in more extreme ways. That culture carried over to Facebook. I've been told that at Facebook recruiters were incentivised with cash to hire more women, but they gave up pretty fast because stuffing the pipeline with weaker candidates didn't work: they got flushed out by the hiring committees.
By the way you're blind to a much more important trick Google uses. Eng hiring was, during my time, at least somewhat well defended against diversity abuse. But firing was totally controlled by HR. And it is an open secret that at Google it's nearly impossible for women to be fired, regardless of how low their performance is. Instead they get moved around between teams. Men, on the other hand, can be and are let go for merely not exceeding expectations. Obviously that's one way to get less male workforce, if you're patient!
1. Assigning the most reliable interviewers to women ensures that women who do meet the bar don't get inaccurately rejected.
2. The hiring bar is enforced at the on-site, not at the phone screen.
3. Sourcing an all-women group from a hiring event and sending them through the normal interview process is ... sending them through the normal interview process.
If you want to claim that Google is trying harder to hire qualified women than qualified men, sure, I don't think anyone disputed that. The dispute was whether Googlers ever think an unqualified person was hired to meet a diversity goal.
(And how often does Google fire someone for discovering that their interview process emitted a false positive? Given the incredible cost of making that mistake, most hiring processes are tuned to err towards false negatives instead of false positives, and Google's intensive process should be particularly good at that. Unless you're claiming HR is firing qualified men to meet a quota - i.e., textbook wrongful termination at scale - a bias for never getting around to firing women can't have a noticeable effect.)
This is a lie perpetuated by googlers. That’s a false dichotomy. There is no evidence that their process reduces false positives. Pointing to a bunch of false negatives just means they have a process that produces a bunch of false negatives. For how smart googlers are purported to be, you would think they wouldn’t report such an obviously illogical statement.
They aren't opposed goals; an organization with a meritocracy that over-optimizes on too few metrics is fragile and collapses in a dynamic market.