Female, minority students took AP computer science in record numbers

This might mean that people whose views are exclusive are excluded from evangelizing those views. The reasons for those are better expressed by Karl Popper's essays on the tolerance of intolerance.

So you can't be against gay marriage or sex before marriage, and be for the suppression of expression. But that's a good thing for the reasons that Karl Popper brings up.

So, I am really not talking about SV startups directly when I talk about the benefits of diversity. I am mostly focused on someone injecting a comment in a design meeting that saves the team weeks of work.

Erm, no.

My points, to avoid confusion:

1) contextual recruitment means you choose the candidate with better potential for success. That's not any more discriminatory than putting a mental weight to a particular university someone studied at, and is the basis for pretty much every "best practice" recruitment process out there.

2) we need white, middle class men, + a load more other people as well. How is that not clear?

Likewise.

> That's a fair example. Today? Likely not. The 1st generation however? I imagine plenty of women/etc felt resentment towards men for being unfairly chosen in the market.

I'd imagine that's true too - however I can't quite understand why we'd know this, and then hold onto the same resentment if the tables are turned more fair now (not even that I'm arguing for it to be anywhere near as explicit as that was, we're talking about tipping factors and increments, not outright discrimination against a group).

> I just mean that, the moment we stop choosing based on criteria that matter to the subject (skill/talent/etc), and instead choose on criteria that matter to a 3rd party subject (skin color/etc)

Again, the argument isn't that we're selecting them based on that 3rd party characteristic though, but using that as a lens to the demonstrated skill/talent. To choose another characteristic - someone who went to an inner city school in a deprived neighbourhood getting a B+ vs someone who went to a private school getting an A-. The former far outperformed their relative expectation, and being within the same ballpark as the latter, would arguably make a better hire. But if we reserved it strictly to looking at "who comes out on top", then the latter could be the only rational choice, even if (given their position) they should have gotten an A+.

> inherently be hiring less qualified individuals

We can argue around this point, but I think it's important to draw it out: hiring someone isn't an absolute. "Qualified" isn't a definitive bar, but a spectrum that can be interpreted in different ways (at least for the vast majority of roles).

> shoehorning more women into X field is really an effective approach

Again, can we move away from implying this is some kind of forcing? We're not dragging underrepresented groups into these roles, but giving them organic opportunities.

I honestly don't buy into the argument around "gender specific" roles - there are enough women who are interested in them already to show that isn't the case. We're literally trying to shift societal expectations - this can't be done in one generation, which is why I respectfully discount an anecdote on that timeframe for this.

We have several hundred years of society deeming that certain roles are for men, and some are for women, and then... men go into male roles and women into female roles. Doesn't mean anything other than we've configured a system and it's behaving as expected, and there's enough exceptions to prove that those societal expectations are worthless, and if they were no longer there, there's the potential for us to be in a much different place (with more, much needed, talent).

> So I think we'd need to be very sure that we're acknowledging the costs now and if they're worth a perceived gain in the future.

I would again like to draw out my point that this isn't about gifting an unsuitable applicant a job over someone who is suitable. There might be costs, but there are also costs to perpetuating the notion that there are some jobs some people aren't supposed to be doing (both explicitly, and implicitly through having no role models doing them).

My closing statement is therefore (based on my point above) - contextual recruitment that gives weight to the performance of underrepresented groups in technology (whilst still ensuring they're technically capable) is a net positive for the company, and the longer term recruitment pipeline.

While your quote mentions class, and asserts that they were not getting signals from class -- one look at the name list chosen for white people will show you how they did not properly control for class. For example they have almost no low class names on the white side. I am sure adding a Bo, Ada, Billy, Bucephelus, Cleatus or Clyde would have for sure changed the callback results on whites.

At the time the study was done I was working at a company that would have never posted a advertisement in a paper, but would have called every single last person both white and black back. The suggestion is that people who posted in papers might have been of a different age group and other means of replying to a job ad such as email at the time may have produced more balanced results.

There is also the issue of the name being disliked and not actually the perceived color of the person. Aversion to unfamiliar ethnic names is entirely different than hating black people. While it is will play out poorly for those with such names it does not guarantee that the same person with a different name would have experienced the same callback rate. This is a phenomena that transcends race. This also is probably a good time to point out assimilation is part of living in a mixed race society. If I recall correctly, the notion of black names started sometime after the 1960s. This happened more so in low-income racially divided areas. It was part of Afrocentrism. While there is nothing wrong with that in general, it can have unintended consequences that are not necessarily racially charged. Meanwhile other immigrants had taken on more traditionally American names, as a result those groups seemed to have done very well in the US (Asians). To keep it short, sometimes it does not pay to be unique.

But none of that matters, because I never claimed there was no discrimination. But I have repeatedly stated how to solve the problem without using more discrimination. That is to pump money and education efforts into low income areas. More aid to single parent families, access to computers (not just access, but actual computers in their names). Classes on personal finance, and dealing with money. Cultivating a notion of delayed gratification. Paid classes for low income parents to learn more about what they can do to make their child successful. Subsidized match savings accounts that are released in full on graduating high school. And 3 free meals at low income schools. I think it would also be interesting to see how to desegregate low income black schools from low income white schools. Don't forget the teachers! Competitive salaries that can support hiring the best of the teachers for these areas.

This is the key because it gives the kids real tools for life that will serve them regardless of how they are treated once they get out of school. A good education also helps with racism in general. If you look the people thinking about these issue and trying things are very educated. If you target poor areas with a relentless education effort both whites and non-whites will be educated, and in return lower the total racial tension, creating a new wave of diversity that will be a lasting one.