>The discrimination faced by Black Americans is because of their skin color, not their socio-economic status.

They're a self-reinforcing loop. A lot of racism is affecting socio-economic status (redlining, no generational wealth) and the bad socio-economic status then fuels the continuation of the disadvantaged status alongside racism.

Is the higher rate at which those with black pigmentation are found guilty of robbery and murder because of their skin color? Or is the skin color only to blame on things we pick and choose?

Using population level pigmentation correlated statistics is an utterly baffling, and confused, way to design a hiring process.

The answer to your question is yes.

They are more likely to be convicted, and are given longer sentences when they are: https://www.ussc.gov/research/research-reports/demographic-d...

They are 7.5 times more likely to be wrongfully convicted: https://deathpenaltyinfo.org/news/report-black-people-7-5-ti...

Having someone convicted of murder (or other felonies) while employed by you is a serious hazard. If we're allowed to use aggregate pigmentation-linked population statistics (which I think is dumb), then it is rational for me to deny candidates because of a fear that my innocent workers are 7.5x more likely to end up convicted and in jail yes?

All else equal I would pick the employee with 1/7th the chance of being wrongfully convicted, as having my employee in jail is bad for business.

This would be a good case for anonymizing resumes in the interview process. I'm not sure why that wasn't considered over setting minimum representation requirements. An easy way to eliminate discrimination is to make it impossible to discriminate between protected classes.

Reducing everyone to a list of past experience and skills would produce knock-on effects such as preferring people with specific existing experience over talented beginners. And in an industry with an extensive, existing problem of discrimination you'd just be maintaining the status quo.

It's also essentially impossible to anonymize resumes in a way that would provide meaningful distinction between people.

Resume design, for instance, is a fairly strong signal for how a developer might think, or their personal attention to detail and craftsmanship.

You'd be amazed how often we don't even need to read a resume to tell whether someone is whether someone is worth hiring.

But people's work history is visible regardless of whether their resume is anonymized. If what you wrote in your previous comment is true, then anonymizing resumes should be strictly better for black applicants. Sure, it won't "solve" disparities in past work experience. But non-anonymous resumes still have that disparity and racial discrimination - anonymous resumes at least eliminate the latter. So if what you wrote about in your previous comment is true, companies should see some increases in hiring rates among black applicants under an anonymous hiring regime.

More broadly, I think you're confusing "non-discriminatory" with "equal chances". A blind audition doesn't mean all participants have equal chances of success. Someone who's been playing the violent for 25 years is probably going to have a better shot than someone who has only a few years of experience - and the lattice has better chances than me, who has zero experience. The fact that we have unequal chances of landing a spot in the orchestra isn't evidence of discrimnation. It's the system working as intended: the more skilled musician has greater chances of getting a spot, regardless of factors like race, gender, etc.

In the context of tech hiring, the purpose of the hiring process is to confer greater chances of success for candidates with relevant skills and ability. If someone is hiring for a position demanding C++ experience, then candidates who are C++ wizards are more likely to succeed at this job opening than people who occasionally dabble in C++.

First, anonymized resumes would not create increased in hiring rates. It might increase the candidate pool through the screen, but behold, racial discrimination is a function at every stage of the interview process.

I am not confusing "non-discriminatory" with "equal chances". I have no idea why anyone would. Obviously, a person with 12 years experience coding in a specific language or stack would be preferable to hire, depending on budget, to someone with none.

That is not the issue any of these measures address, or even should address. No one is arguing we should be hiring people with a GED as physicians because they're "diverse".

You're arguing at a straw man.

With Zoom interviews, it's possible to anonymize candidates throughout every stage of the interview process. Anonymous resumes, anonymous interviews. This would entirely eliminate racial, or gender, discrimination.

That would be true if people didn't sound different.

You can alter the audio signal to mask gender: https://blog.interviewing.io/we-built-voice-modulation-to-ma...

Similar tech could be used to mask accents.

I suspect the real reason people are so hesitant to anonymize interviews is that the disparities will persist - or even grow larger - and it'll be more difficult to ascribe it to racism or sexism.

Accent and pitch are not the only ways in which people sound different.

If technology could be developed which masked pitch, accent, idioms, patterns and approach, I suspect it would change hiring rates.

I suspect the reason you think such technology wouldn’t improve the situation is because you’re a bigot.

> If technology could be developed which masked pitch, accent

This is already possible. Idioms are more specific, but how often do you encounter an identifying idiom in a technical phone interview. "Patterns and approach" are part of the software development skills being measured - it's like saying a blind orchestra discriminates between people of different musical skill.

> I suspect the reason you think such technology wouldn’t improve the situation is because you’re a bigot.

Then let's start using anonymous hiring tools and find out who's right. But again, for some reason, those who most ardently claim that interviews are biased against "diverse" candidates are oddly resistant to anonymization.

> Idioms are more culturally-specific not racially specific. I encounter greater differences in idioms between urban and rural people than along racial lines.

You need to revisit what "culture" means, and how many people of a given race share a culture.

> "Patterns and approach" are part of the software development skills being measured.

This was about speech patterns and conversational approaches, which vary greatly across cultures, both socio-economic, national and internal.

> Did you miss the empirical evidence to the contrary linked earlier?

You posted a blog post, from a startup. It was not peer-reviewed. It was not readily duplicatable. I'm not sure how to have a conversation with a person's whose standard of evidence for "genders and races are inherently inferior" is "a guy on a website said it."

Culture is vastly variable even inside a given race. Heck even within inside a given race, inside a given country culture is highly variable. It's probably much easier to distinguish between a costal urban educated white person and a rural white person, than distinguishing between two people of different races in the same location.

You can easily test whether anonymization is working: have interviewers try to guess the identity characteristics of the applicant and if they're able to distinguish between them then it's failing. If the interviewer is unable to infer it, then the anonymization is working.

> I'm not sure how to have a conversation with a person's whose standard of evidence for "genders and races are inherently inferior" is "a guy on a website said it."

Who on earth said this? This reads like a complete no sequitur. Who is saying that genders and races are inherently inferior?

I'm not sure how to have a conversation who thinks that the mere notion that not all disparities are due to bias means "genders and races are inherently inferior." Men are vastly overrepresented among murder convictions. If someone says this is not because of bias, but because men commit more murders are you going to accuse them of sexism?

> preferring people with specific existing experience over talented beginners.

Well that sure beats institutionalized racism

It just replaces one problem with a different one, or, worse, will mask the actual problem through "facially neutral" policies that are anything but.