Hacker News Reader

cost/performance of the employee

I wonder how good people really are at measuring the latter. In some jobs such as sales, you can get a pretty good approximation - although it's still hard to measure which sales persons build lasting accounts vs. which ones overpromise and then dump the cost of unmet expectations onto technical support or account maintenance people. (Edit: obviously you can measure it in a lot of repetitive type jobs like facotry works etc., but that doesn't seem appropriate for this context.)

It's even trickier in environments like technology. Twitter's a huge company, but how do you measure the productivity of their employees? The stock price is down on what it IPOed for, user acquisition growth has slowed and the firm has never made a profit, so should we conclude that the average Twitter employee has negative productivity? Obviously not, that would be a highly misleading generalization - but it's no more foolish than some of the generalizations I've seen offered to justify executive compensation or equity allocation at some companies.

Economically we'd like to maximize the return on capital invested in hiring someone, but if we don't have a way to measure that objectively then 'performance' can end up as just a bullshit story to rationalize essentially arbitrary management hiring decisions of the staff who fall within 1 or 2 standard deviations of the norm (excluding the infrequent brilliant or burnout employees).

This rests on the assumption that hiring and interviewing is meritocratic. It's not, it's a hugely flawed system. And in the valley there is a HUGE preference towards "fit". So much so that some companies even prioritize "fit" as a primary hiring criterion.

That is a major part of the problem. The solution isn't to start prioritizing "diversity" when hiring folks, the solution is to avoid hiring just within your narrow comfort zone of people who are like you. And that includes people with different life experiences, educational backgrounds, etc. as well as people with different cultural backgrounds and ethnicities. Mostly that's just a matter of becoming a better, less biased, more objective interviewer.

Another factor is acceptance and prejudice. Keep your own prejudices in check and police other folks. When you see or hear of instances of people being insensitive or of excluding someone based on their cultural or other differences, act on it, and help fix it. If you want women to be part of your team, then maybe don't consider having a meeting at a strip club. You'd think that sort of thing is a no brainer, but especially in the SV world full of man children who have discovered there are no rules other than the ones they impose on themselves, it can be a bit of a problem.

Suppose we adopt a more meritocratic system - base hiring decisions mostly on code tests, github accounts, and similar things that are almost entirely based on performance.

(I mostly do this already. When I do a 1 hour interview, I have a ladder of 4 standard questions I ask, followed 1 question about a topic on the person's resume that I understand well. Just to see if they really know what they claim to know.)

Would you be happy with such a process? Note that most SJWs (at least, the pro-female SJWs, who rarely discuss race) tend to oppose this sort of process on the grounds that it will result in fewer women.

> github accounts

Hiring by github/open source contributions mean you're selecting for programmers who have ample free time and an interest joining the overwhelmingly young, white, and male open source community.

You're going to filter out women, black people, and anyone with a family. You're also going to filter out people who are bound by nasty non-competes, or even people who just don't like to program outside work.

It's a tough one: you can clearly see how people navigate communities and get a much more in-depth picture of their development process, but you're also leaving people out that don't fit the mold.

> Would you be happy with such a process?

Until someone figures out how to do a blinded interview, focusing strictly on applicable skills and being conscious of bias is as good as it'll get.

No one is filtered by anything other than their own choices. There is nothing preventing a woman from using github - github doesn't even ask about gender or race.

I use github, I've put hundreds of hours of work into projects on github. But you can't see that because those repos aren't public. Could I, in theory, work on some open source projects on github? Sure. But to be honest the main reason to do so would be job related, to use github as a CV. Personally I think my time is better spent working on the projects that I already have going on (one of which is software dev for a charity that helps sick children) or starting on projects for my own business. There are plenty of good reasons for people to not have tons of contributions to open source projects on github, and ignoring that fact does nothing but artificially restrict your talent pool.

If I see good code, I know that (absent some sort of fraud) you are a good coder. That's a pretty solid "lets talk". In numbers, I'll interview 3 people with good code and maybe hire 1.

If I don't see good code, you put yourself into the same category as all the bad coders out there who can't code. I may need to interview 10 or 20 people before I find someone decent - I simply can't differentiate between you and them without spending hours on code interviews.

Obviously I'm going to talk to the people with visible code first, and only hunt for the needle in a haystack if I get desperate.

I'd point you to a few well written articles that address this line of thinking pretty well:

* http://www.ashedryden.com/blog/the-ethics-of-unpaid-labor-an...

* https://blog.jcoglan.com/2013/11/15/why-github-is-not-your-c...

* http://www.garann.com/dev/2012/you-keep-using-that-word/

Those articles do not address my claim at all.

Ashe Dryden merely lists some personal choices of women (cleaning their house more, taking care of children), and some unrelated statistics about women in general. If she did somehow draw a causal link between those activities and not using github (she doesn't), it would imply that women should also be absent from Facebook, Pinterest and Instagram. They aren't.

The second link is merely an animated gif of an octopuss.

The only point the third article makes is based on the implicit and unproven assumption that merit is uniformly distributed.

None of those articles remotely attempt to show that women are prevented from using github.

> The second link is merely an animated gif of an octopuss.

Now that is funny. Either the blogger changed his URL structure in the time I copy/pasted, or I did a dumb:

https://blog.jcoglan.com/2013/11/15/

> If she did somehow draw a causal link between those activities and not using github (she doesn't), it would imply that women should also be absent from Facebook, Pinterest and Instagram. They aren't.

Wait a minute, are you seriously trying to advance the argument that software development (the purpose of github) is as mentally similar as browsing the web for pictures of cakes (the purpose of Pinterest, as far as I can tell)?

Not sure if serious or trolling.

E: jcoglan is doing referer: madness, copy/paste the URL or bang refresh.

The claim I make is that if a woman is on facebook/pinterest/etc, she clearly has the time/connectivity to use github.

Mentally it is certainly different - but discriminating based on mental differences is entirely the purpose of a hiring filter.

SJW is such a demeaning term that can be applied to anyone who points out something they feel is wrong in the world. So how are you supposed to point out things which are wrong and not "overdo it" like some lame SJW?

Or should we just all be "cool" and not point out the fucked up things people do or say throughout our lives?

I'm really tired of people gas-lighting others who dare to say "Hey, that's not okay" to the co-worker doing the typical "chinese person speaking engrish" impression or joking about how the woman interviewing for the software engineer position probably isn't very good.

What is your stance on this? What turns someone from a decent person into a social justice warrior to you? Am I an SJW in your eyes for this response? Why?

> You no longer are opting to choose based on cost/performance of the employee. are we seriously okay with this? Isn't that demoralizing to the new hire? Now every moment, they are wondering if they got hired because of who their parents are, versus whether they earned it or not.

If you're a white male who is honest with himself, you'll understand that being a white male carries with it a set of privileges.

Just consider: a white child is significantly more likely to go to a school with experienced teachers than a black child, more likely to take honors classes in high school, more likely to get into a good university, and when they graduate, are more likely to get job offers based on nothing more than their white male name[1].

Let's not fool ourselves: humans are not the coldly rational actors we like to think. Ignore the fact that I've already got a leg up on networking because my (white male) peers work the jobs I want to work. When it's time for me to go in for an interview for a software engineering position the odds are overwhelmingly high that every person who interviews me, from the team interview up to senior management, will look a lot like me. That goes a long way in me getting the job - our monkey brains are pretty well wired to look at someone that looks like ourself and think "yeah, he's a competent, trustworthy individual!"

So yeah, I don't think anyone's actually hired in the cold cost/benefit analysis you talk about. There's lots of chances for significant bias to creep in. In reality, if people were hired solely based on the criteria you talk about, and race/gender/... were not considered, I bet we'd see more diverse workforces than we have now.

If someone can figure out how to set up a blinded interview that would be interesting.

[1] http://www.nber.org/papers/w9873.pdf

As a white Eastern European male in a US university, I had a privilege of paying out-of-state tuition out of my pocket (which was about 8x the in-state amount), while being denied most forms of aid and not allowed to work. And I had to take classes about how privileged I am! Didn't feel very privileged. Could it be that the concept is flawed?

Joke answer: what is intersectionality, precious?

Real answer: privilege is attained from an intersection of many factors. How many times have you heard someone say "Sure, I'm a white man, but I grew up on a rural farm to poor parents. Where's my privilege?" And in a way, they are correct - socioeconomic privilege is a real thing, and if you're poor the deck is stacked against you.

But put Tim, the poor white farmer, and John, a rich black man, behind the wheel of a BMW. Which one belongs there? Which one gets stopped by the police? Hand Tim an airplane ticket to fly to NYC, and hand one to Hamza, and see who gets the extra TSA pat-down.

In your case, are you an immigrant or on a student visa? Either way that's a disadvantage for sure. Are you as privileged as Biff Kennedy, who had a Nantucket yacht club membership at age two and graduated Groton as an all-star lacrosse player? Nope. But when you and Biff go in for an interview, management is gonna look more like you two than John or Hamza. (Of course, Biff's father already arranged for him to get the job. And if he hadn't, the connections he made from his Skull and Bones days would be more than enough. But that's just Biff's privilege!)

The concept is only flawed under the most cursory of understanding and analysis.

> The concept is only flawed under the most cursory of understanding and analysis.

I find that many counter-arguments to privilege will try to dial back the level of analysis to make privilege seem ridiculous.

> You no longer are opting to choose based on cost/performance of the employee. are we seriously okay with this?

I think this is a fair question. If diversity is weighted non-zero for hiring, it means some other factor's weight in the hiring decision was reduced. Maybe that's a completely fine trade-off, but it does exist.

> You no longer are opting to choose based on cost/performance of the employee

If performance is actually better with a diverse team then preferring diversity in the hiring process is based on cost/benefit.

> whether they earned it or not

Businesses don't care if you "earned it" they care about cost/performance. My parents are both quite intelligent and up until high school I learned more from them than I did from school. If I had less awesome parents I would have a significantly harder time finding a job. It's not fair, but it does not demoralize me.

So I cannot say I have personally reviewed the studies on diversity improving outcomes, but I have a suspicion that it applies only when all other things are equal (as in talent/skills/performance were all as equivalent as they could be vs the non diverse comparison). My point is there is hardly ever a scenario where you are presented with >1 candidates which have equivalent skills and some happen to be diverse.

Someone else made the point though, that we probably suck as determining level of skill/performance anyways? I can personally attest to that, and maybe based solely on that fact, hiring for diversity might be a winning strategy, I simply don't know. I know I am terrible at judging others skill/performance during an interview, and am striving to improve it.

You can help by attempting to break your coworkers of bad habits. There's enough to joke at about the nature of code and system administration that they don't need to joke about ethnicity and sexism, too.

It does matter how many blacks, whites, women, men, gays or straights you hire, if they are all 20-something graduates of the same 10 CS programmes, you have not achieved any meaningful sort of diversity.