5 ms·

> Imagine google made an error and it turns out that their analysis has a sign error. Blacks/women actually underperform. Would you favor flipping the sign on w

> Imagine google made an error and it turns out that their analysis has a sign error. Blacks/women actually underperform. Would you favor flipping the sign on whatever action you propose? If not, why not? No, because of the above point about diversity. It has an effect on more than just the individual. I don't know if Google has made any claims about the following, but it's been noted that, for example, diverse management teams tend to outperform nondiverse ones. > I admit, I do sometimes provide ideas that no one else has, mainly in the area of bayesian stats, low level performance hacking and probabilistic data structures. Is that what you mean? Are these the secret powers unavailable if you don't have someone of my race? You are being intentionally facetious. I know you're a smart person, that you're very probably aware of what diversity arguments are about, and this reply is thoughtless in a disrespectful way. I'm only spelling it out to avoid downvoters: the idea is that different approaches and attitudes are informed by race, culture, background and gender in ways that are difficult to measure in the interview process. For the record, I would consider a white hillbilly or southerner as 'diverse' in most computer workplaces.

yummyfajitas 10y ago
No, because of the above point about diversity. That point does not apply to lending or crime prediction. There is no collective decisionmaking on the part of parolees or borrowers. Do you have any objection to using your same solutions, but with the opposite direction, in this areas? E.g., if google should try and hire more blacks (due to them overperforming), should BofA try to lend less money to them (because they underperform)? Similarly, since we know diversity harms group cohesion, should companies which value group cohesion avoid it? You are being intentionally facetious. I know you're a smart person, that you're very probably aware of what diversity arguments are about, and this reply is thoughtless in a disrespectful way. I'm only spelling it out to avoid downvoters: the idea is that different approaches and attitudes are informed by race, culture, background and gender in ways that are difficult to measure in the interview process. I don't know why you find this disrespectful. Could you clarify? I'm simply asking, concretely, how "race, culture, background and gender" inform different approaches and attitudes to detecting systematic failures in a sensor network? Or similarly, how does this inform different approaches to querying a dataset consisting of a normalized SQL schema together with a collection of very large structured files? I'm being facetious because I consider this idea to be patently ridiculous. In most of my jobs I've been the only person of my race present, and it's literally never affected any professional task. That's also why I'm asking for specifics, rather than vague allusions to some study you kind of remember reading. I do of course know what the real argument is about. Diversity is a sacred value and all the arguments you've brought up are actually just post-hoc justifications for a behavior you virtuous. I'm just arguing that there is little more than that.
- cynicalkane 10y ago
  Well, I'll give you an example. It's popular among computer nerds to dismiss other peoples' ideas. They might do something like attack someone's ideas with intentional misunderstanding, lazy thinking, and facetious reasoning. In short, they do not show the bare minimum of consideration for others' ideas, instead looking first for avenues of attack. Having ignored and attacked the other person, they then claim no disrespect was meant. They might say something like "I don't know why you find this disrespectful. Could you clarify?" It does not occur to them that being able to attack peoples' ideas in this pompous way is actually a position of privilege. Anyone can do it, but people born a step ahead are more likely to not reflect on their position, or even feel entitled to it. In my experience, women and people from disadvantaged backgrounds--people who were not praised to the moon for writing "Hello World" in eighth grade--are less likely to behave in this way. This is self-evidently an enormous advantage on engineering teams. Usual disclaimer: personal experience, your mileage may vary, this is just one example, privileged backgrounds have advantages too, yada yada yada.
  - yummyfajitas 10y ago
    I did not attack you or ignore you. I directly addressed one of your arguments and attempted to determine whether the other argument is your "true rejection" [1]. On the contrary, the only person ignoring the other's points and making anything close to an attack is you. You've accused me of being disrespectful, called me lazy and accused me of intentional misunderstanding, while simultaneously NOT addressing any of my core arguments. Quite entertaining, I must admit. [1] Specifically, I think even if that argument could be proven false, it would not change your conclusion at all. So bringing it up is merely a distraction. See also: http://lesswrong.com/lw/wj/is_that_your_true_rejection/ http://lesswrong.com/lw/wj/is_that_your_true_rejection/
    - cynicalkane 10y ago
      The question was if diversity is valuable in engineering. I gave a specific example of a specific kind of diversity that is valuable to engineering teams. If you're interested in Google's data, they make an excellent search engine that you can use to find that kind of stuff. I encourage you to do your own research, in part because I think I've done enough work leading you to the proverbial water.

yummyfajitas 10y ago

No, because of the above point about diversity. That point does not apply to lending or crime prediction. There is no collective decisionmaking on the part of parolees or borrowers. Do you have any objection to using your same solutions, but with the opposite direction, in this areas? E.g., if google should try and hire more blacks (due to them overperforming), should BofA try to lend less money to them (because they underperform)? Similarly, since we know diversity harms group cohesion, should companies which value group cohesion avoid it? You are being intentionally facetious. I know you're a smart person, that you're very probably aware of what diversity arguments are about, and this reply is thoughtless in a disrespectful way. I'm only spelling it out to avoid downvoters: the idea is that different approaches and attitudes are informed by race, culture, background and gender in ways that are difficult to measure in the interview process. I don't know why you find this disrespectful. Could you clarify? I'm simply asking, concretely, how "race, culture, background and gender" inform different approaches and attitudes to detecting systematic failures in a sensor network? Or similarly, how does this inform different approaches to querying a dataset consisting of a normalized SQL schema together with a collection of very large structured files? I'm being facetious because I consider this idea to be patently ridiculous. In most of my jobs I've been the only person of my race present, and it's literally never affected any professional task. That's also why I'm asking for specifics, rather than vague allusions to some study you kind of remember reading. I do of course know what the real argument is about. Diversity is a sacred value and all the arguments you've brought up are actually just post-hoc justifications for a behavior you virtuous. I'm just arguing that there is little more than that.

cynicalkane 10y ago
Well, I'll give you an example. It's popular among computer nerds to dismiss other peoples' ideas. They might do something like attack someone's ideas with intentional misunderstanding, lazy thinking, and facetious reasoning. In short, they do not show the bare minimum of consideration for others' ideas, instead looking first for avenues of attack. Having ignored and attacked the other person, they then claim no disrespect was meant. They might say something like "I don't know why you find this disrespectful. Could you clarify?" It does not occur to them that being able to attack peoples' ideas in this pompous way is actually a position of privilege. Anyone can do it, but people born a step ahead are more likely to not reflect on their position, or even feel entitled to it. In my experience, women and people from disadvantaged backgrounds--people who were not praised to the moon for writing "Hello World" in eighth grade--are less likely to behave in this way. This is self-evidently an enormous advantage on engineering teams. Usual disclaimer: personal experience, your mileage may vary, this is just one example, privileged backgrounds have advantages too, yada yada yada.
- yummyfajitas 10y ago
  I did not attack you or ignore you. I directly addressed one of your arguments and attempted to determine whether the other argument is your "true rejection" [1]. On the contrary, the only person ignoring the other's points and making anything close to an attack is you. You've accused me of being disrespectful, called me lazy and accused me of intentional misunderstanding, while simultaneously NOT addressing any of my core arguments. Quite entertaining, I must admit. [1] Specifically, I think even if that argument could be proven false, it would not change your conclusion at all. So bringing it up is merely a distraction. See also: http://lesswrong.com/lw/wj/is_that_your_true_rejection/ http://lesswrong.com/lw/wj/is_that_your_true_rejection/
  - cynicalkane 10y ago
    The question was if diversity is valuable in engineering. I gave a specific example of a specific kind of diversity that is valuable to engineering teams. If you're interested in Google's data, they make an excellent search engine that you can use to find that kind of stuff. I encourage you to do your own research, in part because I think I've done enough work leading you to the proverbial water.