Uh, yes? I'm saying if we somehow magically determined that group A had 10% benefit than group B at a task, then we should expect to see more of group A represented, even if their populations are equal.
Exactly like we expect to see certain phenotypes over-represented in basketball.
Despite the difficulty, saying it's hard but it just must be true that everyone's equal doesn't seem like a robust approach. Or is Google saying it just has zero way of evaluating a person's work? Perhaps they could do blinded code reviews or something.
I'm saying discrimination, meaning non-equal results, is fine. I don't think there should be any active discrimination, and incorrect unintentional discrimination should also be removed.
I'm saying discrimination, meaning non-equal results
I think you may want to reconsider the word you use for that, because that isn't a widely shared definition.
To be clear, when you say "discrimination.. [snip] is fine" most people hear that as "making hiring or promotion decisions based sorely or primarily on the gender of the person is fine". That's what the dictionary definition is too.
This isn't discrimination, and (almost) no one says it is. Unequal outcomes have many sources, and provided they are non-discriminatory that's defensible.
The issues arise when unequal outcomes occur and no one can agree on the cause, because there is a (very good) chance that there is an inequality of opportunity somewhere in the process.
I suspect there is a miscommunication here and your defense of the possibility of unequal outcomes is being confused by the wording.
I think the other person with which you are speaking is communicating clearly enough without your help on what the definition of words are supposed to mean.
Exactly like we expect to see certain phenotypes over-represented in basketball.
Despite the difficulty, saying it's hard but it just must be true that everyone's equal doesn't seem like a robust approach. Or is Google saying it just has zero way of evaluating a person's work? Perhaps they could do blinded code reviews or something.