Should he try to have more tact? Sure, of course. But failing that, the correct response is compassion and gentle correction, not a witchhunt. Not driving him out of his jobs.
In particular, he says he did try to do the compassion and gentle correction work, and Stallman had no interest in it.
Stallman's job has a job description - leading the Free Software Foundation. If he's not leading it effectively, it is important for his own cause that he step away, and it is important that the FSF Board do their job by finding someone more effective at it. If he's not suited for it for reasons outside of his control and never will be, that's all the more reason that we should compassionately and gently get him to find another job.
(Also, I think this argument is deeply unfair to the many neuroatypical people in this world who still manage to not say the things Stallman does - which is in most cases not merely a matter of "tact," but even so, many of whom put conscious, learned effort into having tact. I have friends with all sorts of things going on in their brain who are wonderful people and in many cases also wonderful free software authors, and I think this argument does them a disservice.)
I'd agree with Matthew Garrett here. Some of Richard Stallman's statements in the past have bordered on being outrightly indefensible for various reasons. Not necessarily indefensible as the opinions of an individual, as controversial as they may be, but certainly inappropriate as the opinions of a person representing an organisation. I doubt that Stallman's various eccentricities have done much to hamper the FSF's mission, but Stallman himself would certainly do little to endear it to anyone.
Edit +1 hour: this comment certainly aged well...
So the (implicit) argument "This is just what you expect with neuroatypical people" / "If we want to be welcoming of neuroatypical people, we have to be welcoming of people who act like RMS" is even less true because of how diverse the category is. We know RMS isn't being excluded because of his (hypothetical) neuroatypicality because there are so many neuroatypical people who aren't and wouldn't be excluded.
If you want to argue that the problem with RMS is he is in the narrow category of people who act like RMS (and I would say that not even Linus Torvalds is in this category), that's a fine argument to make, but it's a bit circular to defend his inclusion on such grounds. Is this category the sort of category, like race or class or neurological makeup, that is beyond (or mostly beyond) one's control and has little correlation with whether you can do good work? Or is this a category like "asshole"?
I'll grant that some of the headlines about what RMS wrote were overblown, but still. You don't have to look outside the four corners of what he wrote to see why it wasn't an excusable thing to blast out to a departmental email. Once you expand from those four corners to his prior comments on pedophilia it gets way worse.
At a certain point, if you have lived as long as RMS has, you should have some knowledge of your own limitations. It would take an absolute master of rhetoric to make his argument and not have it taken badly. He is not one, and he should (at this point) have the barest humility to not subject the entire CSAIL mailing list to his half-baked ideas on how Minsky might be exonerated.
I also think that being tactless is not a moral crime. It's aesthetically offensive and counterproductive from a leadership perspective, but being a bad leader doesn't make somebody a bad person.
And it’s deeply unfair to people who have issues to simply blame gross behavior in being “neuro-atypical”.
Sorry but you don’t get to weasel out by saying you have poor social skills.
He's attempting to find some remotely plausible scenario where the actions of his friend are morally acceptable.
No amount of tact makes that okay, and complete lack of tact just makes it worse.
