The section on overlapping bell curves is disappointing.
> if you pull out a random man and a random woman, you have about a 77% chance the man will be taller
> you’d have about a 2% advantage over sheer random chance if you bet on the man to be more invested in achieving
I have no doubt that math is correct, but that's not the point. A small difference in the means becomes a big difference at the tails. And Google does select for the tail.
There are a few sex-linked neuroscience differences linked to innate rather than learned behavior, but they tend to be very low-level skills which aren’t directly linked to high level tasks, much less performance at a job which requires many high-level skills. A slight difference in, say, verbal versus spatial memory doesn’t tell you who’s a better programmer, and in studies it’s consistently been found that top performers mix a variety of skills.
Google employs on the order of 140k people and it’s unlikely that more than a few of those are so dependent on one of those narrow skills that it would dominate hiring, much less that the bias would always run in a single direction.
> if you pull out a random man and a random woman, you have about a 77% chance the man will be taller
> you’d have about a 2% advantage over sheer random chance if you bet on the man to be more invested in achieving
I have no doubt that math is correct, but that's not the point. A small difference in the means becomes a big difference at the tails. And Google does select for the tail.
This response did the math: https://medium.com/@ryanbaldini/worth-noting-small-differenc...