The first is that I feel that text is an AWFUL medium for communication for these sorts of discussions and criticisms, unless expertly crafted (as this article was, I feel). Numerous times I've had arguments with friends via text that get out of hand because of subtle misinterpretations of intent and phrasing, where as in person they're much more civilised. You can draw a few conclusions here: either we suck at communicating via text and need to get a lot better, or discussions of this nature are better had face to face, where intent and nuance of expression are easier (for many) to judge. What feels like normal discourse and 'clinical' (for lack of a better word) discussion to you may not feel the same way to others, simply because short-form text is an imperfect medium.
The other point of contention I have is that it's fine to treat diversity programs with scientific rigour, but you also need to allow for the proper timescales in which to study the effects. If you subscribe to an intersectional feminist viewpoint, then it stands to reason that you need to observe the success or failure of such diversity programs over a generational timescale in order to assess their efficacy. Change on that scale simply isn't going to be visible over a one or two year period before you decide the diversity program doesn't work and you axe it.