I understand your argument but it's taking a very narrow view of best. Specifically, you're usually making a very small selection from a small sample that responded to a specific job description.
Now the chances are that you haven't selected the best is some global absolute sense as that would be stunningly unlikely. So you've actually selected from a sample that is exceptionally biased in numerous ways. By having a drive for, say, female developers, you're just replacing one very biased sample with another very biased sample.
Now, if you're at a sufficiently small organisation and you're recruiting sufficiently few people, and you're only interested in the immediate needs of your organisation, it is reasonable to assume that a female biased cohort won't necessarily lead to better candidates due to the current heavy gender imbalance.
However, not everyone recruits with such constraints. For example, I work for a large, global organisation that recruits 100's of devs a year into a pool of 1000's of devs. Given the statistical significance, any gender imbalance is not because we recruited the best, it's because we recruited the best of a very badly skewed pool. The corollary is that, if we don't try to address this, we are deliberately ignoring a huge pool of potentially talented devs.
Now, it would be correct to argue that, if we were just to select women from the existing pool, it isn't likely to help things much. If an organisation is doing that then they're mostly doing it to appear diverse. But a drive for female devs that affects the pool absolutely can make a difference. It's not easy, it can mean planning for the long term i.e. going to universities, schools etc. but I strongly believe that it's worth it.
Now the chances are that you haven't selected the best is some global absolute sense as that would be stunningly unlikely. So you've actually selected from a sample that is exceptionally biased in numerous ways. By having a drive for, say, female developers, you're just replacing one very biased sample with another very biased sample.
Now, if you're at a sufficiently small organisation and you're recruiting sufficiently few people, and you're only interested in the immediate needs of your organisation, it is reasonable to assume that a female biased cohort won't necessarily lead to better candidates due to the current heavy gender imbalance.
However, not everyone recruits with such constraints. For example, I work for a large, global organisation that recruits 100's of devs a year into a pool of 1000's of devs. Given the statistical significance, any gender imbalance is not because we recruited the best, it's because we recruited the best of a very badly skewed pool. The corollary is that, if we don't try to address this, we are deliberately ignoring a huge pool of potentially talented devs.
Now, it would be correct to argue that, if we were just to select women from the existing pool, it isn't likely to help things much. If an organisation is doing that then they're mostly doing it to appear diverse. But a drive for female devs that affects the pool absolutely can make a difference. It's not easy, it can mean planning for the long term i.e. going to universities, schools etc. but I strongly believe that it's worth it.