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.
Right, annoying that Slashdot has no way of editing posts. Fortunately there are few situations where non-knowledge sabotages knowledge (as it would if it were AND).
When the doctor decides whether you are competent to make that decision, they do give an official answer. If a healthy and normal person demanded it, they would probably say he was irrationally depressed.
So they make a value judgment based on the contents of your life, on how much pain you are in, on what your prospects are. etc. Some lives are deemed rational to want to end, others are deemed not rational.
Such as system does not aim to answer the question if a life is worth living
If it tells some people that they are irrational in demanding to die (e.g saying they are depressed), and others that they are rational in demanding to die, then it implicitly does.
So, a carefully designed and managed system is capable of determining whether your life is not worth living? Presumably they will also find that some people are wrong in wanting to die, otherwise they wouldn't need a system at all.
Which lives are worth living or not sounds to me like the kind of question it's maybe not right to set an official answer to.
I have sympathy with people who feel life isn't worth living. But I wish they would not demand that others validate their choice by killing them.
If you don't comply with a lawful order, you're taking a pretty big risk.
Remember the thing they say about loans? If you owe the bank $100 that's your problem. If you owe the bank $100 million, that's the bank's problem. Microsoft, Google, Oracle, they're pretty damn important to the US economy. Also, unrestricted executive power isn't entirely unopposed in government. If
The point of this is also that these orders, even if "lawful", give government a headache about how to enforce compliance. You can't just send Tim Cook to jail without an explanation. If Jeff Dean mysteriously turns up dead in a ditch somewhere, lots of people are going to get cold feet, and start asking themselves whether they want to be part of this anymore. The silicon valley elite have a lot of power to say no, more than they probably admit to themselves. Yes, there are risks, but it's not nearly as bad for them as less powerful people (e.g. Ladar Levison) face.
I agree with your point, but Jim Henson was also an writer, and if you ask me a highly original and talented one. The actual performance with the puppets was a small thing of what he made.
The spice girls weren't the only or the first, but one of the most clearly and well-executed "music group" concoctions (cynically) aimed at a specific demographic.
Oh, they weren't remotely the first. A really old example is the Comedian Harmonists, formed in 1928. It was assembled after a newspaper ad calling for auditions, and the plan was explicitly to create a German equivalent of a foreign group that had impressed the originator, namely The Revellers. That was exactly how Boyzone was formed too, only there the foreign group was Take That.
Village People was also formed after a newspaper ad and auditions. There, the purpose was maybe not so much to make a localized version, as a subculture-inspired twist on an established concept.
That's how Spice Girls were formed too, as an obvious twist on the well-established concept of a boy band. But they didn't succeed in the market they were formed to target, which was teenage boys! Luckily for them, they were surprised by the emerging market of pre-teenage girls.
All these bands were formed after ads and auditions, but some of them - perhaps the most deserving of being called "boy bands" - were formed by managers or other people who weren't themselves performers.
A way to give them a headache, is just to quietly not comply. Not post the national security letter, not shout anything from the rooftops, just refuse to do as you're told. They either have to punish you openly, drawing attention to what they were trying to do, or go full criminal and "cause something bad to happen to you" (which has risks of its own for them).
I have argued that once several, actually important people started to do this (e.g. silicon valley tech sector employees, or even execs), the government would quickly find out the headaches weren't worth it, and change policy - either to the better, or at least to more visible aggression. Don't underestimate the power of sand in the machinery.
This is par for the course for chess programs of that generation, though. I remember an old DOS chess program that would allow "castling" by just switching the king and rook, regardless of whether there were pieces between them. One Win 3.11 chess program I remember, was ridiculously easy to break if you used the take back a move function. Some sort of internal data structure wasn't properly cleaned if you did that, so the computer would start making illegal moves, randomly putting pieces on the board, etc.
Slashdot Mirror
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.
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.
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.
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.
Right, annoying that Slashdot has no way of editing posts. Fortunately there are few situations where non-knowledge sabotages knowledge (as it would if it were AND).
Not on the argument that the people studying them are paid, no. Then you have to dismiss regular medicine as well.
I have a skill level of 15. My skill is represented by this string:
000000101111100010001101101011001
You have a skill level of 16. It is represented by this string:
111100001110010100100111001001001
Our team's skill is represented by this string:
(bitwise AND of the two above)
111100101111110110101111101011001
As you can see, we have a combined skill level of 23.
This is a very basic answer to how having a diverse team can help.
The anti-science crowd's go-to way to dismiss any science you don't like.
When the doctor decides whether you are competent to make that decision, they do give an official answer. If a healthy and normal person demanded it, they would probably say he was irrationally depressed.
So they make a value judgment based on the contents of your life, on how much pain you are in, on what your prospects are. etc. Some lives are deemed rational to want to end, others are deemed not rational.
If it tells some people that they are irrational in demanding to die (e.g saying they are depressed), and others that they are rational in demanding to die, then it implicitly does.
Involving other people so they become complicit in your death doesn't sound like demanding nothing to me.
If you know how some people use suicide threats (and even suicide attempts), that "maybe a little compassion" doesn't sound quite so innocent.
So, a carefully designed and managed system is capable of determining whether your life is not worth living? Presumably they will also find that some people are wrong in wanting to die, otherwise they wouldn't need a system at all.
Which lives are worth living or not sounds to me like the kind of question it's maybe not right to set an official answer to.
I have sympathy with people who feel life isn't worth living. But I wish they would not demand that others validate their choice by killing them.
Remember the thing they say about loans? If you owe the bank $100 that's your problem. If you owe the bank $100 million, that's the bank's problem. Microsoft, Google, Oracle, they're pretty damn important to the US economy. Also, unrestricted executive power isn't entirely unopposed in government. If
The point of this is also that these orders, even if "lawful", give government a headache about how to enforce compliance. You can't just send Tim Cook to jail without an explanation. If Jeff Dean mysteriously turns up dead in a ditch somewhere, lots of people are going to get cold feet, and start asking themselves whether they want to be part of this anymore. The silicon valley elite have a lot of power to say no, more than they probably admit to themselves. Yes, there are risks, but it's not nearly as bad for them as less powerful people (e.g. Ladar Levison) face.
I agree with your point, but Jim Henson was also an writer, and if you ask me a highly original and talented one. The actual performance with the puppets was a small thing of what he made.
Oh, the vast majority of music scores and poetry would benefit from the makers thinking of it more as skill, and less as art,
Oh, they weren't remotely the first. A really old example is the Comedian Harmonists, formed in 1928. It was assembled after a newspaper ad calling for auditions, and the plan was explicitly to create a German equivalent of a foreign group that had impressed the originator, namely The Revellers. That was exactly how Boyzone was formed too, only there the foreign group was Take That.
Village People was also formed after a newspaper ad and auditions. There, the purpose was maybe not so much to make a localized version, as a subculture-inspired twist on an established concept.
That's how Spice Girls were formed too, as an obvious twist on the well-established concept of a boy band. But they didn't succeed in the market they were formed to target, which was teenage boys! Luckily for them, they were surprised by the emerging market of pre-teenage girls.
All these bands were formed after ads and auditions, but some of them - perhaps the most deserving of being called "boy bands" - were formed by managers or other people who weren't themselves performers.
A way to give them a headache, is just to quietly not comply. Not post the national security letter, not shout anything from the rooftops, just refuse to do as you're told. They either have to punish you openly, drawing attention to what they were trying to do, or go full criminal and "cause something bad to happen to you" (which has risks of its own for them).
I have argued that once several, actually important people started to do this (e.g. silicon valley tech sector employees, or even execs), the government would quickly find out the headaches weren't worth it, and change policy - either to the better, or at least to more visible aggression. Don't underestimate the power of sand in the machinery.
This is par for the course for chess programs of that generation, though. I remember an old DOS chess program that would allow "castling" by just switching the king and rook, regardless of whether there were pieces between them. One Win 3.11 chess program I remember, was ridiculously easy to break if you used the take back a move function. Some sort of internal data structure wasn't properly cleaned if you did that, so the computer would start making illegal moves, randomly putting pieces on the board, etc.