Alright, I'm going to take another swing at this. I will reorder things.
> Says the pot to the kettle.
That was my point. When I said "it's like an accent" I didn't mean I myself don't have an accent. It's that I don't perceive it. People do not perceive their own accents as accents. They don't perceive their own biases as biases.
But accents are still accents regardless of whether we perceive them. Biases are still biases.
My point wasn't that I am pure, free of sin. It wasn't about me at all. It was that biases are bad but often invisible.
> They wanted to have a diverse workforce, and came up with an excuse for it post-hoc. The best defense I've seen of this is that diverse opinions are good for business.
Is it true? If so, who cares whether it was post hoc?
I don't know whether you've done any machine learning. Nowadays all the news is about LLMs, but back in the day there were other algorithms -- decisions trees, support vector machines, logistic regression models, simple Bayesian models, etc. You had a mess of data. You defined feature vectors. You vectorized your data and threw it at the algorithm. You got a classifier. Or a regression model or whatever. Then there were also meta algorithms that took these base algorithms and combined them to make something more robust and accurate -- a random forest or something. This meta algorithm worked only if these sub-deciders differed. You make one decision tree, duplicate it 1000 times, and wrap it in a random forest and you get nothing but wasted effort. To achieve the gain you needed your deciders to have different opinions.
This is the theory behind diversity in an organization: you have multiple viewpoints plus some mechanism to combine them into a final decision. It has nothing to do with race or gender or anything. It is a provable way to get better decisions.
Did someone also want to lift up people historically beaten down? Maybe? Thats also a good thing! It's awesome if you can get better decisions and also make a more just society, right?
> We have a word for this: racism.
What? This is just a shibboleth. Is racism a good thing or a bad thing? Is trying to counteract racism just as bad as raw racism? It produces a more fair outcome, but it's forbidden! We must preserve the injustice because to counteract it would also be unjust!
This is a ludicrous position. The point of using a DEI mechanism is that it is just that, a mechanism. You set it up and let it decide so that flawed human judgment doesn't decide. Is the mechanism racist? I don't know, does it have ideas and intentions?
Perhaps it will be clearer if I try to do some perspective taking for you.
Imagine there is a company with a South Asian CEO, CTO, etc. They only trust South Asians. Only South Asians are a good cultural fit. North Americans are lazy, unmotivated, dumb, incurious, entitled. They are poor team players. They won't put in extra hours, or if they do, it will achieve nothing. And then if you interact with them there's always this tension, awkwardness. You can't tell the same stupid American jokes. It's a joke! They have no sense of humor.
Now you are a North American with skills. You actually are very talented and insightful. You have lots of energy, lots of ideas. You would be a great employee. But this company will not hire you. (You disgust them a little.) Is this racism? I've just inverted identities, but I think you would agree that the company is racist and this is unfair. And it's clear that this racism harms both you and the company. But if the company's HR or whatever recognized the possibility that racism was tainting their hiring practices and implemented DEI targets as an impartial mechanism to work against this bias, that would be racist against South Asians! So to maintain their moral purity they should preserve their original racism and deny you the job. [slaps dust off hands and calls it a day]
Do you see the other perspective now? To summarize: racism is bad because it harms both parties. A mechanism that counteracts this helps both parties, and is just and good to boot. Saying that they are equivalent because they both consider race, to say that this makes them both equally racist, is to make "racism" a nullity, a thing of no consequence. If this is racism, it is good racism, because it rights an injustice. Or, alternatively, it is not racism. Racism is using race to commit an injustice, not to counteract it.
> Says the pot to the kettle.
That was my point. When I said "it's like an accent" I didn't mean I myself don't have an accent. It's that I don't perceive it. People do not perceive their own accents as accents. They don't perceive their own biases as biases.
But accents are still accents regardless of whether we perceive them. Biases are still biases.
My point wasn't that I am pure, free of sin. It wasn't about me at all. It was that biases are bad but often invisible.
> They wanted to have a diverse workforce, and came up with an excuse for it post-hoc. The best defense I've seen of this is that diverse opinions are good for business.
Is it true? If so, who cares whether it was post hoc?
I don't know whether you've done any machine learning. Nowadays all the news is about LLMs, but back in the day there were other algorithms -- decisions trees, support vector machines, logistic regression models, simple Bayesian models, etc. You had a mess of data. You defined feature vectors. You vectorized your data and threw it at the algorithm. You got a classifier. Or a regression model or whatever. Then there were also meta algorithms that took these base algorithms and combined them to make something more robust and accurate -- a random forest or something. This meta algorithm worked only if these sub-deciders differed. You make one decision tree, duplicate it 1000 times, and wrap it in a random forest and you get nothing but wasted effort. To achieve the gain you needed your deciders to have different opinions.
This is the theory behind diversity in an organization: you have multiple viewpoints plus some mechanism to combine them into a final decision. It has nothing to do with race or gender or anything. It is a provable way to get better decisions.
Did someone also want to lift up people historically beaten down? Maybe? Thats also a good thing! It's awesome if you can get better decisions and also make a more just society, right?
> We have a word for this: racism.
What? This is just a shibboleth. Is racism a good thing or a bad thing? Is trying to counteract racism just as bad as raw racism? It produces a more fair outcome, but it's forbidden! We must preserve the injustice because to counteract it would also be unjust!
This is a ludicrous position. The point of using a DEI mechanism is that it is just that, a mechanism. You set it up and let it decide so that flawed human judgment doesn't decide. Is the mechanism racist? I don't know, does it have ideas and intentions?
Perhaps it will be clearer if I try to do some perspective taking for you.
Imagine there is a company with a South Asian CEO, CTO, etc. They only trust South Asians. Only South Asians are a good cultural fit. North Americans are lazy, unmotivated, dumb, incurious, entitled. They are poor team players. They won't put in extra hours, or if they do, it will achieve nothing. And then if you interact with them there's always this tension, awkwardness. You can't tell the same stupid American jokes. It's a joke! They have no sense of humor.
Now you are a North American with skills. You actually are very talented and insightful. You have lots of energy, lots of ideas. You would be a great employee. But this company will not hire you. (You disgust them a little.) Is this racism? I've just inverted identities, but I think you would agree that the company is racist and this is unfair. And it's clear that this racism harms both you and the company. But if the company's HR or whatever recognized the possibility that racism was tainting their hiring practices and implemented DEI targets as an impartial mechanism to work against this bias, that would be racist against South Asians! So to maintain their moral purity they should preserve their original racism and deny you the job. [slaps dust off hands and calls it a day]
Do you see the other perspective now? To summarize: racism is bad because it harms both parties. A mechanism that counteracts this helps both parties, and is just and good to boot. Saying that they are equivalent because they both consider race, to say that this makes them both equally racist, is to make "racism" a nullity, a thing of no consequence. If this is racism, it is good racism, because it rights an injustice. Or, alternatively, it is not racism. Racism is using race to commit an injustice, not to counteract it.