4 ms·

Very true. It is a vague and meaningless concept. It has never been defined except by referring to antiquated notions of “race” stemming primarily from 18th an

Very true. It is a vague and meaningless concept. It has never been defined except by referring to antiquated notions of “race” stemming primarily from 18th and 19th century ideas. In modern times, we’ve apparently added biological sex, sexual preference and a vague (and meaningless) concept of self-identity that can, apparently, cross all other definitions with impunity - even changing fluidly throughout the day. To appoint people as “head” of this mess of ideas point to deeper issues in the organization.

danShumway 5y ago
The fact that categories are often fuzzy does not make them useless as a general heuristic for finding people who are more likely to share similar experiences with each other. Almost every kind of categorization system shares the problems you talk about: people can call themselves "programmers" for a variety of different reasons, even though they're working in wildly different environments and might be totally unsuited for each other's jobs. People self-identify with political parties, even though there are no strict rules about what a "Democrat" or "Republican" are, to the point where beliefs of members of those parties might overlap with each other or even be the opposite of what the party normally espouses. We separate games/movies into genres, even though genres are often a complete mess that don't always map well to player preferences or experiences. We take a light spectrum and we break it up into distinct colors, even though the number of colors and lines between those colors are completely arbitrary and meaningless. It's good to recognize that categories are fuzzy human inventions that are prone to abuse, and it's good to be cautious about putting a lot of weight on them in every situation. It's good not to assume that everyone in a category is identical. But abandoning categories entirely would make it much harder (even impossible) to go through normal day-to-day life. So it is similarly useful to be able to think about categories like "Black", "Gay", "Conservative", "Christian", "Rural", "Trans" -- even as we acknowledge that those categories are just human inventions and that they blur around the edges. Those categories are still be a useful heuristic for trying to find people who are more likely to share certain attributes or who have had certain life experiences.
- slibhb 5y ago
  > Those categories are still be a useful heuristic for trying to find people who are more likely to share certain attributes or who have had certain life experiences. In other words, prejudice is a "useful heuristic".
  - danShumway 5y ago
    > In other words, prejudice is a "useful heuristic". If you don't see any difference at all between community and othering, or if you think we need a single universal metric that defines whether categories are being used for evil in every single scenario, then I don't think you understood what I was saying about fuzziness. Categories are a neutral tool, it's how they're used and the effects of their usage that determines whether or not an action involving them is problematic -- just like every single other tool in existence.
- mikaeluman 5y ago
  You are trying to rationalize the usage of “black” and “white” while simultaneously seeming to agree with my point. I would say this to you: if you believe a diverse set of skills, experience and personality are important (I believe they can be) then use categories for that. Skills categories are judged from education and practical work tests. Experience is judged from CV and references from earlier employments. Personality can be assessed from tests and judged during interviews. These aren’t perfect either, but they are far more relevant than: “oh, you seem brownish, how nice”
  - danShumway 5y ago
    Multiple issues: - First, this is not an either/or situation. You can assess personality, skills, and experience in traditional ways, while simultaneously understanding that having a diverse collection of genders, races, ages, disabilities, etc on a team is often a competitive strength. You don't have to throw away people's CVs to do any of this. There's an assumption that I see sometimes that caring about diversity during hiring necessarily means ignoring everything else, and that's just not what people like me are saying to do at all. - Second, the role of a head of diversity should go way beyond simple hiring decisions, it can also mean trying to make a more welcoming environment that accommodates diverse workforces after they're hired, it can mean trying to explain why certain demographic outcomes and trends exist in hiring, and trying to figure out on a high level if there are areas where Google is biasing its hiring process or culture to be less diverse. It should be kind of obvious, but Kamau wasn't just sorting applications based on skin color -- you wouldn't need multiple people and a "head" to do that. A diversity initiative is supposed to look at the entire company, identify subtle problems in multiple areas/departments, and come up with creative solutions to those problems. - Third, I think you're discounting life experience in general as a contributing factor towards how people make decisions. It's not "you're brownish, that's nice." It's "Black people want to use our products, is there someone in the room who's going to notice if our AI mislabels them in photos?" It's "we want our UI to be usable by blind people, it would be nice if we had anyone in our office that can identify certain pitfalls." It's "we're designing a social network feature that will accidentally make doxing much easier, what are the odds someone who's LGBT+ in our office will raise their hand and point out how dangerous this is for their community?" Those aren't theoretical concerns, they're issues that companies like Google care a lot about for very practical reasons. So it's a big jump to act like political diversity, age, life experiences don't impact how products get made. Understanding diversity means both acknowledging that these communities are fuzzy and that not everyone in them is identical, and also acknowledging that on net people in these communities have experiences and perspectives that can be valuable to incorporate into decision making processes. - Fourth, you're missing that many diversity initiatives in companies don't exist in a vacuum. Google has a diversity initiative because it has problems with diversity, because there are unreasonable biases in its hiring processes and recruitment strategies against older, POC, and other minority applicants. Diversity initiatives in companies like this don't just randomly appear out of nowhere. They're not trying to get rid of good hiring practices, they're a response to existing flaws in the hiring process. It's all well and good to argue that hiring processes should be egalitarian and that diversity will naturally arise if we ignore the categories. But the current outcomes aren't egalitarian, and ignoring the categories didn't lead to natural diversity. Argue if you want that the kind of diversity I'm talking about is irrelevant to quality, but doing so really only reinforces that the lack of diversity across the tech industry is abnormal and it's worth having teams inside of companies who try to figure out why their hiring outcomes seem to be biased in certain directions.

danShumway 5y ago

The fact that categories are often fuzzy does not make them useless as a general heuristic for finding people who are more likely to share similar experiences with each other. Almost every kind of categorization system shares the problems you talk about: people can call themselves "programmers" for a variety of different reasons, even though they're working in wildly different environments and might be totally unsuited for each other's jobs. People self-identify with political parties, even though there are no strict rules about what a "Democrat" or "Republican" are, to the point where beliefs of members of those parties might overlap with each other or even be the opposite of what the party normally espouses. We separate games/movies into genres, even though genres are often a complete mess that don't always map well to player preferences or experiences. We take a light spectrum and we break it up into distinct colors, even though the number of colors and lines between those colors are completely arbitrary and meaningless. It's good to recognize that categories are fuzzy human inventions that are prone to abuse, and it's good to be cautious about putting a lot of weight on them in every situation. It's good not to assume that everyone in a category is identical. But abandoning categories entirely would make it much harder (even impossible) to go through normal day-to-day life. So it is similarly useful to be able to think about categories like "Black", "Gay", "Conservative", "Christian", "Rural", "Trans" -- even as we acknowledge that those categories are just human inventions and that they blur around the edges. Those categories are still be a useful heuristic for trying to find people who are more likely to share certain attributes or who have had certain life experiences.

slibhb 5y ago
> Those categories are still be a useful heuristic for trying to find people who are more likely to share certain attributes or who have had certain life experiences. In other words, prejudice is a "useful heuristic".
- danShumway 5y ago
  > In other words, prejudice is a "useful heuristic". If you don't see any difference at all between community and othering, or if you think we need a single universal metric that defines whether categories are being used for evil in every single scenario, then I don't think you understood what I was saying about fuzziness. Categories are a neutral tool, it's how they're used and the effects of their usage that determines whether or not an action involving them is problematic -- just like every single other tool in existence.
mikaeluman 5y ago
You are trying to rationalize the usage of “black” and “white” while simultaneously seeming to agree with my point. I would say this to you: if you believe a diverse set of skills, experience and personality are important (I believe they can be) then use categories for that. Skills categories are judged from education and practical work tests. Experience is judged from CV and references from earlier employments. Personality can be assessed from tests and judged during interviews. These aren’t perfect either, but they are far more relevant than: “oh, you seem brownish, how nice”
- danShumway 5y ago
  Multiple issues: - First, this is not an either/or situation. You can assess personality, skills, and experience in traditional ways, while simultaneously understanding that having a diverse collection of genders, races, ages, disabilities, etc on a team is often a competitive strength. You don't have to throw away people's CVs to do any of this. There's an assumption that I see sometimes that caring about diversity during hiring necessarily means ignoring everything else, and that's just not what people like me are saying to do at all. - Second, the role of a head of diversity should go way beyond simple hiring decisions, it can also mean trying to make a more welcoming environment that accommodates diverse workforces after they're hired, it can mean trying to explain why certain demographic outcomes and trends exist in hiring, and trying to figure out on a high level if there are areas where Google is biasing its hiring process or culture to be less diverse. It should be kind of obvious, but Kamau wasn't just sorting applications based on skin color -- you wouldn't need multiple people and a "head" to do that. A diversity initiative is supposed to look at the entire company, identify subtle problems in multiple areas/departments, and come up with creative solutions to those problems. - Third, I think you're discounting life experience in general as a contributing factor towards how people make decisions. It's not "you're brownish, that's nice." It's "Black people want to use our products, is there someone in the room who's going to notice if our AI mislabels them in photos?" It's "we want our UI to be usable by blind people, it would be nice if we had anyone in our office that can identify certain pitfalls." It's "we're designing a social network feature that will accidentally make doxing much easier, what are the odds someone who's LGBT+ in our office will raise their hand and point out how dangerous this is for their community?" Those aren't theoretical concerns, they're issues that companies like Google care a lot about for very practical reasons. So it's a big jump to act like political diversity, age, life experiences don't impact how products get made. Understanding diversity means both acknowledging that these communities are fuzzy and that not everyone in them is identical, and also acknowledging that on net people in these communities have experiences and perspectives that can be valuable to incorporate into decision making processes. - Fourth, you're missing that many diversity initiatives in companies don't exist in a vacuum. Google has a diversity initiative because it has problems with diversity, because there are unreasonable biases in its hiring processes and recruitment strategies against older, POC, and other minority applicants. Diversity initiatives in companies like this don't just randomly appear out of nowhere. They're not trying to get rid of good hiring practices, they're a response to existing flaws in the hiring process. It's all well and good to argue that hiring processes should be egalitarian and that diversity will naturally arise if we ignore the categories. But the current outcomes aren't egalitarian, and ignoring the categories didn't lead to natural diversity. Argue if you want that the kind of diversity I'm talking about is irrelevant to quality, but doing so really only reinforces that the lack of diversity across the tech industry is abnormal and it's worth having teams inside of companies who try to figure out why their hiring outcomes seem to be biased in certain directions.