What kind of "females" are you hanging out with? Elderly members of the British Royal Family?

No one is approaching a random group of stranger "females" to tell dick jokes. There just happens to exist somewhere, in public, a place with lot of dick jokes. People going in there, men or women, are idiots if they think it's supposed to make them feel uncomfortable. If it does make them feel uncomfortable anyway, well, tough luck! The world is full of things that would make us uncomfortable if we sought them out. Hope you never have to deal with someone from another culture, if you can't even handle some crude humor from your own.

Sadly, 985 of them suck ass, which makes this a meaningless statistic.

No, the statistics are still valuable, unless you make an argument that Windows has a higher percentage of shitty games. Absent any other information, it's reasonable to assume the percentage of awful games are similar on all platforms.

(But in fact, I think that the existence of more shovelware-friendly middleware on Windows means Windows has a higher percentage of bad games).

Well yeah, technically it's Valve that helps people get their games running on multiple platforms, not Stream. But that's still picking nits.

That is the wrong question to ask. The right question is "why should we not blindly the believe the guy who says the bug isn't in his code?" and the answer ought to be obvious.

Well, Google's fact engine has a long way to go. If they start downranking thing that aren't "facts", Wikipedia just got a lot more power... I wanted to find out how high Denmark's highest hill is, and Google gave me a number, but it also said it was a giant flaming mountain of doom. They had that from the wikipedia page, of course, but they still presented it in their factbox.

So I expect Google will start downranking flaming mountain of doom-denialists.

It does not claim (as I understand it) to represent every scenario, merely a special case of a specific scenario.

Freeman Dyson wouldn't be bombastic and exaggerate, would he? "Prisoner's dilemma has been solved!"

For actually intelligent strategies (and the point of strategies is that they should be intelligent), the folk theorem is the relevant solution, not this. For that matter, this seems like a weak, specific case of the folk theorem.

Which means that we haven't been hiring the best and the brightest. We've been hiring those who are similar to us.

I agree, and I totally support blinding of resumes, and as much blinding as possible in general. But there's one thing you overlook, which leaves your argument vulnerable:

It's possible that hiring the best and the brightest is not the wisest move. It's possible that it's a good idea to hire those who are similar to us. I don't think so, but it's possible, we don't know.

To take a concrete example, take the study that showed lab assistants were rated more poorly with a female name on the resume. That's solid proof of gender prejudice. But playing the devil's advocate here, we don't know that it's unjustified prejudice. Perhaps the people evaluating the resumes have had tons of lab assistants of both genders, and found a clear tendency that the women performed worse than their resumes suggested, and the men performed better.

Thing is, even if that were true, I'd advocate for blinding. It's not for efficiency's sake that we should end discrimination, but for justice's sake: You didn't choose to be born a woman, you deserve to be judged on individual merits. Even if women on average were awful at this job, that information should be off-limits to use in hiring decisions, because using it would be a great injustice to those who are not awful.

This is of course even more salient in the case of race and ethnicity. Because while it is highly implausible that women should be worse lab assistants, we do have crime statistics, and if people were allowed to discriminate based on those, it's quite possible that a shop owner could "reasonably" deny Roma entrance to his shop, for instance. It will probably reduce shoplifting! But it's also a horrible injustice to those Roma who do not shoplift. It doesn't matter if that is 90%, or 10%. It doesn't matter if there's just one honest Roma in the whole country. No individual should answer for the statistical proclivities of a category he didn't choose to be in and can't even escape.

But this also shows why blinding yourself to information about race and gender can't just be a "best practice". That asshole shop owner who denies Roma entrance to his shop, he's doing a great injustice, but he might well a comparative advantage over more fair shop owners. Being just can be costly, and because of that, it's important that we demand sharing that burden fairly. We can hope that when we do, we find that it isn't so costly after all, maybe it's even a net benefit. But we must never base our demand for justice on such hopes. Justice first, then profit.

Intel knows that an investment in diversity is insurance against the government.

It's not government they're afraid of, it's their rivals' PR teams, and powerful people in media who on a whim might decide to throw their power around.

Yes, it could be, in fact. Few companies are on such an existential knife-edge that they can't afford to make a few godawful decisions. If you don't believe that, I have a couple of teambuilding activities and motivational speakers to sell you.

As it happens, I think that companies do rush to hire competent women, and even less competent women due to quite rational reasons (a company's productivity is not simply the sum of its employees skills). It's a supply problem, and it starts long before high school. But whether I am right or wrong about that, an argument that a business sector can't possibly collectively and systematically make poor decisions, is a weak argument.

"Judgment" is a weasel word here. In one sense what you say is true, but not if you read that word like most people read it.

It impairs judgment in the sense that it slows your thinking. A slow brain still comes to the same conclusions as a fast brain most of the time - it just takes a little longer.

If it is the sort of judgment that improves with more time thinking about it, then alcohol impairs judgment. But e.g. whether it's a good idea to go for a nighttime swim in the canal, or whether it's a good idea to hit on your boss' wife, are not especially time-dependent judgments. If alcohol makes you do stuff like that, it's social conditioning at work, not brain impairment.

Yes, conditioned response is a factor too, but still that's a conditioned response based on social conditioning, not the biochemical properties of alcohol. Here, anthropological studies are useful to take a look at. Alcohol does not universally reduce inhibitions. In some cultures, they even split up the beliefs about alcohol, so that e.g. liquor from the city makes you aggressive, boisterous and disinhibited, but traditional fermented beverages just makes you calm and mellow.

Something rather important is that all or virtually all of that effect is in your socialization and expectations around alcohol, not the alcohol itself. There are plenty of classic studies showing that people who believe they consume alcohol, behave as if they really did - and conversely, that alcohol does very little to your inhibitions unless you figure out that's what they're feeding you.

So no, it doesn't really make you do things you normally wouldn't do. It just gives you an excuse - one your surroundings believe in, and one you probably believe in yourself.

If we didn't have alcohol, I bet that either we would find something else and ascribe inhibition-reducing properties to it, or we would act slightly less inhibited all week instead of just concentrated to friday night.

The socializing isn't a property of the alcohol. Sure, in a society where everyone drinks alcohol and you are seen as an outcast and weird if you don't drink alcohol, then not drinking means less socialization and less happiness, possibly less lifespan. But if you ask me, the blame for that should be placed on the culture, not the people who refuse to conform to it.

But yes, as Ben Goldacre pointed out long ago, in the UK at least most people drink, and those who don't are probably different in lots of other ways that potentially affect health. That was always one reason to be suspicious of the health benefit claims.

There were others. The argument for health benefits of moderate alcohol consumption was always based on the weakest forms of EBM with blinders on: no randomized controlled trials, and rarely any theories about what would cause the mysterious health benefits at low levels. All which were presented were either discredited, e.g. antioxidants in red wine - or not really a very good reason to recommend drinking when you come down to it (the social benefits you mention).

The OP asked about diversity, not just any specific kind of diversity. But either way, you rarely know beforehand just what sort of diversity will be useful, and just knowing that people are different might affect your decisions.

Suppose you have three new employees, and are sending two of them to work in a team. You have tested them, and you estimate they all have skill level 16 individually (estimating exactly the bit string, exactly which areas are their strong and weak points, your testing isn't good enough to do). Now two of the employees are men, and one is a woman.

In that case, why should you pick the woman and one of the men? Because men and women are slightly different - meaning that you have reason to think they may have slightly different strengths and weaknesses. There might be ever so slightly fewer "collisions" in the bit string in a man/woman pairing compared to a man/man pairing.