And once again, the characterization of "allowing the debate" means one thing for James (why, he "cited studies") and another for everyone else (they are an angry mob). The most telling bias in this piece is that characterization.
Perhaps is James had not hamfistedly "cited" population research (as Brooks suggests) but then given very specific personal-level fixes (e.g., pair programming , suggestions of "pipeline" fixes, etc) he would not have cast quite so much doubt over his intent.
What's also lost in this summary is one of the most important points: long term exposure to stereotypes has a powerful influence on people (many references of varying quality here: https://apcentral.collegeboard.org/courses/resources/women-m...). By embracing them, we actually create self-fulfilling prophecies.
These prophecies may be based on a statistical mean, but what's lost in that simple numerical distillation is what harm befalls even modest outliers to the distribution. Stereotypes which may seem obvious and unimportant to 3/4 of a population may be a crushing burden and source of relentless stress to the remaining quarter.
It's interesting how many of my peers fought to liberate themselves from stereotypes of "weakness" and "inferiority" that were tied around them as smart teenagers. But when it comes time to recognize the harm in these stereotypes to outliers in a other group, they appeal to the same logic that oppressed them. One might argue that these traits are adopted defense mechanisms well-impressed by abuse. I'm not sure that justifies them, though.