For instance, in some places (you may be aware of ;) there is a societal bias that heterosexual anglo men are hyper sexual. That they are literally untrustworthy due to it is not an uncommon bias. It is an easy jump from that bias to an implicit (or even explicit) belief that it is better NOT to hire engineers to build streaming event architectures who are constantly thinking about sex and therefore to build a pipeline that biases against heterosexual white men. A company that did this, intentionally or not, is at a disadvantage when hiring because they have removed from the applicant pool people for the wrong reasons.
A company that instead was blind to that bias (or actively worked against it) would therefore have a wider pool to select from.
Even 1 counter example is often enough to change the biases for everyone who encounters them. So the company that hired 'yummyfajitas' and realized either A) he is not hyper sexual or B) he is but it doesn't impact his ability to perform the task is more likely to detach the wrong biases from their hiring pipeline and this has a multiplying effect.
Substitute "women aren't as good at math", "older people learn technologies slower", "black guys don't research as well" or whatever other bias you want.
Also, you know, sometimes we should do things because they are the right thing to do.
I'm hiring a data team. Based on the applicant pool it'll probably be me and 3 Punjabi males assuming I hire purely based on ability [1]. My claim is that I'll get no marginal value out of hiring a Maharashtrian, an African or a westerner over an equally qualified Punjabi. I provide no special western perspective on Kafka.
I also don't know why it would be the "right thing to do" to hire a Maharashtrian relative over a Punjabi - but that might be because I don't believe in group rights. If you do, I'd love to know how you choose which groups are relevant - there are, after all, 2^{6 billion} subgroups of humanity.
[1] I take pains to do exactly that - I skip resumes and look at github, follow scripted interviews with objective criteria, etc.
I'm a fairly loud, direct person. I work with other loud and direct people. We added someone to our team who is a bit quieter and quite frankly thoughtful onto the team. The effect wasn't so much that (s)he changed our technical decisions, but that we as a team started to give people apace for less confrontational disagreement. It improved our individual abilities to communicate and approach a problem from multiple angles.
That is why diversity in personality is important. The hypothesis is then that race/gender/etc is correlated with intellectual diversity.
Anecdotal evidence yes, but something I've found valuable.
No, it's not. Which is my point.
If you want to make the argument that racial/gender/etc groups think differently, be my guest - but that also implies that an unbiased hiring process might have disparate impact. E.g., if onion2k's hypothesis that women are more emotional than men is correct, then a hiring process aiming for rationalism might hire fewer women.
If there were a systematic bias against westerners that meant they were unlikely to apply for, get callbacks for, or be able to work in your environment with Kafka, actively removing those impediments is an opportunity. The larger the population being biased against and the more acute the need for applicants is, the larger the opportunity.
Framing this as a "hire one person over another" or as a "rights" issue is incorrect. Many of the systematic biases can be corrected for in ways that either don't cost the unbiased set anything, or provide advantages to that set as well. Further, this is an issue of courtesy and empathy, not rights. No one has the right to my courtesy or empathy, but providing it broadly is the right thing to do.
I don't see how the work environment is made better by my presence. I'm not going to be any nicer to minorities than my majority coworkers. And quite honestly, if someone of my race considers the environment better because they have a firangi to talk to, I'm unlikely to get along well with them.
For instance, sick child care disproportionally falls on women. So more women (of course not all women) view flexible hours, flexible sick time and flexible work from home options as major benefits. In many instances providing those things is a net improvement for everyone (including the business productivity) but it isn't as obvious in cases that don't experience the problem (like places that don't have as many parents generally or women specifically).
I suppose improving this environmental factor can be viewed as a PR move (which frequently has a pejorative tone, but I assume you are not applying that) but it also can have intrinsic benefits even if it never causes someone to think differently about the company.
In fact, heterogeneous preferences w.r.t in-kind benefits are an argument for workplace homogeneity (and heterogeneity across workplaces). Women who want flexible hours can go work for a company providing flexible hours. Dudebro neckbeards can work for a company providing unlimited redbull, videogames and vodka even after normal working hours. If they work at the same company someone's pay is reduced to cover benefits they don't use.
Your position on this implies that a) we all know our preference around in-kind benefits b) our use of in-kind benefits is static, predetermined, or correlated job moves c) the provision of in kind benefits is zero sum and/or d) in-kind benefit decisions can be optimally designed.
My position is the opposite on all of these cases. There are many in-kind benefits that I didn't realize were great until I received them because someone else wanted them, the utility of these benefits varies over time and is not correlated from employee to employee (and the homogenity required to correlate them would be hugely detrimental in other resource allocation ways), and that the optimal benefit strategy is often emergent rather than designed, and that the more inputs into the system we have the better the emergent behavior is.
Finally, being polite and showing empathy would also follow your rules about zero sum behavior and workplace homogeneity. I'm curious what that implies about your position on general civility and empathetic behavior in the workplace.
...bringing in ideas that are different to what we have now...In the case of tech, that's women, ethnic minorities, LGBQT+, non-tech, etc.
I wasn't responding to this:
...bringing in people who are superficially different but think the same way in order to manipulate the media...
Obviously PR matters and I'm happy to let my company exploit my race for that purpose (though I probably wouldn't have taken the job if I thought that was a significant factor). But that's completely tangential to the original point.
This sort of stereotype probably has the predictive value of a D20 roll.
Phil's an 11, I spit in his lunch. Nancy is a 19, no problems expected.
Having a more diverse team that includes someone with a more emotionally-driven perspective could lead to someone asking the question "Is the way we're manipulating user behavior actually ethical?"
This isn't about finding different technical solutions to problems; it's about finding different problems.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3108582/
http://journals.plos.org/plosone/article?id=10.1371/journal....
And most emotionally driven ethical discussions on HN lean towards popularity contests - compare the outcry against online harassment of that gamergate girl to the online harassment of Charles Carreon (https://en.wikipedia.org/wiki/Charles_Carreon), for example.
But even if you are right, what does diversity in the traditional sense of your post above have to do with it? Are you suggesting that women/LGBTQ+ people are more emotional and less rational?
Geez, I really hope that that was just an unfortunate turn of phrase. You can't possibly mean that non-(white-males) have a quote "more emotionally-driven perspective", do you?
There is a big difference and yes, "manipulating the user" is sinister.
That's not a lie or bending the truth. It is manipulating the user.
It sounds like you are decorating the showroom to boost sales... I hope it works out well for you. I was thown by the initial desciption that sure didnt sound like faces and buttons: "I'm streaming realtime events and then running statistical models on the datastream in order to manipulate user behavior."
On several occasions I've joined a company and been the only diversity carrier. I can't think of a single advantage beyond PR (I make group photos look more diverse, yay!) that my race has given the company. Any such advantage is miniscule in comparison to the advantages gained by my knowledge of statistical significance, multiple comparisons, regression and the like.
Realistically, if you wanted to add diversity of thought, you'd be far better off hiring a Frequentist of the majority race than a Bayesian of a different one. And if you wanted real (harmful) diversity of thought, you'd hire a magical thinker rather than someone persuaded by evidence.