[centimeter]

Anyone who claims that hiring racism/sexism/ageism/etc. dominates in some industry should be ignored unless they’ve already conditioned on obviously-relevant factors with sex/race/age/etc. correlates, like the desirable strength/intelligence/etc. distribution of workers in that industry.

You might naively claim that professional basketball has a racist preference for black men, until you remember that black men are vastly more likely to fall into the height and athleticism distribution optimal for playing basketball. There are also industries where you rationally expect e.g. whites and asians to be disproportionately represented.

[doorstar]

Are you flat out saying that some races and sexes are not intelligent enough to be in the tech industry?

[runawaybottle]

You’re just asking for a long drawn out fight over which cultures emphasize education more.

[doorstar]

Scratch out race.

Are you flat out saying that some sexes are not intelligent enough to be in the tech industry?

[centimeter]

"Some sexes" is a weird way of saying that, and no. However, women experimentally have a narrower IQ distribution, which means that low and high IQs tend to be dominated by men. It's (I suspect) why there are more male nobel laureates (and programmers) but also more male prisoners.

This is also the expected result if you're familiar with GMV as mediated by sex chromosome pseudodominance in most mammals, including humans.

[doorstar]

> women experimentally have a narrower IQ distribution

A glance at Wikipedia shows that this statement is contested. Pretty much every statement about men, women, intelligence, and IQ is contested.

> This is also the expected result if you're familiar with GMV

OK, I'll bite.

https://bsd.biomedcentral.com/articles/10.1186/s13293-019-02...

"few sex differences (if any) remain statistically significant"

It amuses me that any article about racism and sexism in the tech inevitably devolves to "It's not racism and sexism if we just think that other races and sexes don't perform well in the tech industry."

[centimeter]

> Pretty much every statement about men, women, intelligence, and IQ is contested.

Well obviously, but most of the contests against hard IQ data don’t have much merit.

> <link to paper about gray matter volume>

GMV isn’t “gray matter volume”, it’s “greater male variability” - hence the mention of sex chromosome pseudodominance.

> in the tech industry.

It’s not just the tech industry - every industry will have some selective pressures.

[runawaybottle]

That’s not the reason man. Socially women were driven away from certain fields, that’s all.

[runawaybottle]

I’m not the OP, but I doubt anyones saying that. If anything you’re asking for a long drawn out fight over why some sexes don’t major in STEM more.

The argument is going to boil down to the “pipeline” problem, ‘there’s not enough to begin with for it to be represented in proportion’, which leads to the moral hazard dilemma of do we just started filling quotas.

Edited

[doorstar]

The OP has clarified that he thinks that women inherently do not have the skills needed for the tech industry. It's not an unusual attitude, and challenges your assessment that this is a 'pipeline' problem.

As long as tech workers think that some races have 'cultural' problems and as long as some tech workings think that there are differences between men's and women's brains that make women less suited for tech work, I think we have to stop dismissing this as a pipeline problem.

The prejudice is real and all over this thread posters are happily justifying it.

[centimeter]

You should try to learn how to discern the difference between “<group> doesn’t have the skills for <activity>” and “<group> is statistically less likely to match selective criteria of <activity>”. It’s a pretty critical distinction.

[user982]

They also said that non-white races look too alike for facial recognition: https://news.ycombinator.com/item?id=23462568

[zozbot234]

They didn't say "they all look alike", they said that white faces have more contrast so are easier to differentiate even with a naïve CV algorithm. It's one more source of systemic bias in the ML literature, especially given the comparative scarcity of source data from non-majority groups.

[user982]

> It's one more source of systemic bias in the ML literature, especially given the comparative scarcity of source data from non-majority groups.

centimeter's posts in that thread directly reject your proffered line of thought:

  > your training data does not have enough people with dark skin or African American face features
  This isn’t the issue - the issue is lower variance across black faces in any basis.

[centimeter]

Here's what I said: "if you partitioned the faces by race the output of the SVD would be much wider for white people [...] white people have more light/dark contrast, more hair colors, more eye colors, etc.". This is an obvious fact that is widely recognized by CV practitioners.

[centimeter]

No, but I am saying that there are differences in population-level intelligence distributions across groups. The data is pretty clear on this.

[doorstar]

I'm not sure I understand the distinction. If there are differences in 'population-level' intelligence, then some populations are less intelligent, correct?

If some populations are less intelligent, then it is OK to not hire them in the tech industry.

In your opinion, how do you tell the less intelligent populations from the more intelligent populations?

[centimeter]

If you use population-level characteristics to make blanket individual-level determinations, you are an idiot.

[zozbot234]

You can't use statements about populations to draw inferences about individuals - and we hire individuals, not populations. Your entire framing of this issue is 100% backwards.

[mekoka]

I love seeing comments like this surface on HN, because they serve to show how some people can think of themselves as rational while being so oblivious about their ignorance. It can go on silently for years, because they think it's just politically incorrect to say it, but that the "facts" are true.

This is why, despite what you may believe, "race" issues are still of actuality in SV. You would be amazed at what some of the supposedly smart and powerful people in these circles believe when it comes to this stuff.

[centimeter]

You should consider the hypothesis that the supposedly smart people actually are smart (as one might expect), and have put more thought into this than you.

[mekoka]

I stand by my assertion. Anyone trying to use a concept as misunderstood as IQ to explain a disparity in tech regarding black people is most probably misguided and jumping to conclusion. The other person responding to my comment (and who seems to be in the same boat as you) said: where they're still in disagreement is what causes it and the definition of "race". There's a very good reason for that. There are many other attributes that correlate _very strongly_ with what people like to call "race", that are likelier candidates (based on strong evidence) as factors of IQ disparity.

IQ researchers know that unless you can trace the direct origin of their findings to genes (as it's been done for some diseases), you cannot claim causation. Especially when there is also strong evidence of socio-cultural factors (environment, history, culture, nutrition, education, lifestyle, means, etc). Even the plasticity of IQ suggests that since it's not a number set in stone, it wouldn't be prudent to make wild genetic claims. Unfortunately the public at large doesn't want to hear that part, and would rather run with its own version of what I like to call "tldr science", that purports to explain why the world works the way it does with cliff notes.

In a discussion about bias against a specific genetic group, if you want to bring arguments that excuse the status quo, by basically saying that the discriminated group is just genetically at a disadvantage, you need to damn well know what you're talking about. But as it's common with many "bottom-line" numbers like IQ, people tend to look at them and make the same mistake of drawing overly simplistic conclusions to very complicated and entangled issues.

The issue at hand here is the relative invisibility of black people in the tech scene. You brought up a point. I will assume that you're in tech, as I am. Despite what we pride ourselves to be able to accomplish, nothing in that field places it beyond the grasp of 99% of humans with a functioning brain, whether those individuals belong to a low scoring group or not. It's a challenging field, yes, but you don't need genius level IQ. What you need is a favorable environment (education, food, etc) that adequately prepares you for it. And this is true for most other fields.

The fact that you ran with "IQ" to explain the phenomenon (and that you had many advocates coming to your rescue in that endeavor) is very interesting to me, as it really is symbolic of how otherwise smart people may convince themselves that they've figured something that they really don't understand that well.

As I said earlier, this is how we remain in that situation.

[lopmotr]

Psychology researchers don't have any major disagreement about the relationship between race and IQ. It's well established that blacks have lower IQ than whites. Where they're still in disagreement is what causes it and the definition of "race".

It's amazing to see so many seemingly intelligent people being such avid science deniers. Just because you've demonized the people who believe something different from you doesn't mean they're wrong.

[TeaDrunk]

This rhetoric in commenting on venture capital and tech implies that whites and asians are more likely to fall into some biological distributions optimal for being in tech or being a venture capitalist.

could you explain what they are? Genuine question.

[exolymph]

It's a matter of whether or not you believe IQ test results.

(Personally, I don't have a position, not being familiar with the literature brandished by either side of the debate. I am, however, aware of the debate's existence — hence this clarifying comment.)

[TeaDrunk]

IQ tests aren’t biological the way height is, thus my confusion. Additionally, we can’t know that tech/VC have significantly higher IQs. My inquiry is genuinely asking if this persons rhetoric had any merit whatsoever since on its face it looks dumb

[diffrinse]

Well, at least we got over phrenology... right?

[zozbot234]

IQ results are meaningful, but there's nothing 'biological' about them. Variation in IQ is heavily influenced by all sorts of social factors, and variation across social groups even comes with difficulties in measurement. The meme that 'IQ is biological' is just that, a baseless urban legend.

[SamReidHughes]

Don't be ridiculous. Performance on IQ tests is highly heritable and is also severely affected by metabolic disorders, childhood iodine deficiency, and sleep deprivation. That's biological.

[centimeter]

It's funny when people are so uniformly wrong but write with such confident language. Most estimates based on twin and sibling studies place 60-80% of variance in intelligence as genetically heritable, with a single-digit percentage of adult IQ attributable to known environmental factors.

[zozbot234]

Twins-vs.-siblings studies are irrelevant to variation across large social groups. There's no way anyone can "write with confident language" about any of that variation being biological in nature, given that the social milieus are so starkly different.

[zo1]

Twin vs sibling studies aim at quantifying what effect genetics has on IQ rather than environmental conditions. If there are hereditary components to it, then they most certainly can be mapped to some extent across population groups, as have a lot of other "traits". There are only so many ways to slice this, but it is unfortunately a delicate topic.

[SamReidHughes]

You don't need IQ test results. The SAT measures actual performance at solving math problems.

https://collegereadiness.collegeboard.org/pdf/sat-percentile...

[asdff]

SAT measures your ability to complete the SAT, not your ability to solve math problems. If you are nervous in a testing environment or need to take your time with math problems or write them out beyond the scraps of paper they give you, you will have a bad time on the SAT. Even when I took the GRE for grad school, I ran out of time on the math section. It wasn't difficult, it was just trigonometry and very basic algebra, but I ran out of time anyway. I still got into grad school, so it turned out none of this testing stress mattered at all.