←back to thread

791 points 317070 | 4 comments | | HN request time: 0.722s | source
Show context
tptacek ◴[] No.15009988[source]
Some of the reasoning in this post is very weak.

It's not very long, and its kernel is an anecdote about how her son is interested in programming and her daughter in photoshop. My daughter is also more interested in art than my son (who is more interested in video games). Both would make exceptional programmers, and both have a latent interest. Both are setting a course for STEM careers, but, like all 18 and 16 year olds --- let alone 9 and 7 year olds --- neither has any clue what they're really going to end up doing.

The piece culminates in a recommendation that we focus our diversity efforts on college admissions and earlier stages in the pipeline. But that's a cop-out. We should work on all stages of the pipeline. It's unsurprising that a Google engineer would believe that gender balance can't be addressed without fixing the college pipeline, but the fact is that virtually none of the software engineering we do in the industry --- very much including most of the work done at Google --- requires a college degree in the first place.

Most importantly, though, the only contribution this post makes to the discussion is to add "I'm a woman and I agree with one side of the debate" to the mix. Everything in it is a restatement of an argument that has been made, forcefully and loudly, already. Frankly: who cares?

Edit: I added "some of the" to the beginning of the comment, not because I believe that, but because I concede that there are arguments in the post that can't be dispatched with a single paragraph in a message board comment (through clearly there are some that can.)

replies(35): >>15010018 #>>15010095 #>>15010098 #>>15010099 #>>15010101 #>>15010105 #>>15010129 #>>15010150 #>>15010173 #>>15010194 #>>15010204 #>>15010230 #>>15010247 #>>15010273 #>>15010330 #>>15010345 #>>15010384 #>>15010389 #>>15010415 #>>15010436 #>>15010457 #>>15010460 #>>15010497 #>>15010501 #>>15010518 #>>15010541 #>>15010655 #>>15010665 #>>15011059 #>>15011368 #>>15011653 #>>15012315 #>>15013242 #>>15013891 #>>15015706 #
jasode ◴[] No.15010098[source]
>, and its kernel is an anecdote about how her son is interested in programming and her daughter in photoshop.

Fascinating how different readers take away different salient points. For me, her main buildup was hiring women to meet a "diversity goal" resulted in pressures to hire some women who couldn't do the work. This creates a perverse feedback loop that unfairly taints future women candidates who could do the work -- which ends up undermining the whole point of diversity. Imo, the biological stuff about her son and daughter is more of a side note.

To restate her text, we could say that yes, there are talented female computer scientists like Grace Hopper and NASA's Margeret Hamilton.[1][2] However, if companies lower the bar to hire women who are not competent like them (because diversity is valued over skills), it will inadvertently make it harder to hire future Grace Hoppers and Margeret Hamiltons.

I'm not agreeing or disagreeing with her but her Google observation is getting lost in her boy/girl preferences sidebar.

[1] https://en.wikipedia.org/wiki/Grace_Hopper

[2] https://en.wikipedia.org/wiki/Margaret_Hamilton_(scientist)

replies(5): >>15010179 #>>15010750 #>>15010934 #>>15010945 #>>15012021 #
tptacek ◴[] No.15010179[source]
Because the piece isn't very well written†, it's unclear whether the "infinite loop" of mediocrity she's referring to is something she actually observed, or something she surmises is possible. She is clear and specific, presenting numbers, when talking about things she was personally involved in.

Since the cycle of increasing mediocrity has a prominent float in the parade of horribles conjured by the "anti-diversity" (for lack of any better term) side of this debate, I'm left assuming she didn't see that occur. But she could also clear that up easily.

Finally, an obvious point: evaluation of the performance of an individual software developer is one of the great unsolved problems of software engineering. Virtually all performance evaluation done today is at root subjective. Subjective performance evaluations are easily tainted by prejudice; in fact, you have to work hard not to taint them.

If you think that's different at Google, re-evaluate: Google also runs one of the most famously capricious hiring programs in the industry. Despite constant rituals and genuflection towards data-driven decision making, Google continues to thrive based on its status as a premiere destination for new software developers, despite running a hiring process renowned for the quality of the people it has alienated. There is ample evidence of Google having scaled broken processes.

99% of what I write isn't well-written either, in case this sounds like a jab at the author, who I am not familiar with.

replies(4): >>15010433 #>>15010906 #>>15012563 #>>15012811 #
cargo8 ◴[] No.15010433[source]
I think that her "infinite loop" theory was her perception of the anecdote around hiring at her own startup. It does seem conceivable, but yeah obviously data is light here.

Good point on how the whole interview process is a crapshoot anyway. Hadn't really thought about that aspect, but obviously is a huge opportunity for subliminal bias since it is how it is.

replies(2): >>15010491 #>>15010932 #
1. tptacek ◴[] No.15010491[source]
Another weird attribute of her theory about setting the bar and paying premiums for talent is that her engineering team, according to LinkedIn, is in India --- which is a radically different market for software talent than SFBA. In particular, the gender distribution of CS grads and programmers in India is very different from what it is here.

One possible interpretation --- and there are probably equally credible others --- is that this founder found it difficult to compete in the (overheated) SFBA market for talent, irrespective of gender.

replies(2): >>15012585 #>>15012861 #
2. goldbeck ◴[] No.15012585[source]
Oh interesting. Where did you find this? (My cursory check didn't show turn up any info on that)
replies(1): >>15013233 #
3. mcguire ◴[] No.15012861[source]
That makes it even more weird. The percentage of women programmers there is higher.

https://m.cacm.acm.org/magazines/2015/5/186026-decoding-femi...

4. tptacek ◴[] No.15013233[source]
I used the following advanced sleuthing techniques (don't share outside HN):

1. I went to LinkedIn

2. I searched for "Silverlabs"

3. I found the page for their company

4. I clicked on the link labeled "14 employees and former employees have LinkedIn profiles"

5. I observed that all the engineering profiles were in Hyderabad.

I'm being a little snarky but also it's good to know how superficial this "research" was so it's not at all unlikely that I'm totally wrong about this.