A "joke" in the glibc manual

Because a project doesn't communicate - individuals communicate on behalf of a project. Possibly several individuals will team together to create a communication on behalf of the project.
In such cases, it would be entirely consistent with the apparent purpose of the Guidelines for those individuals to use the guidelines in determining how they, personally, will communicate.

I don't think it is consistent with the stated purpose of those guidelines to use them to judge how other people are communicating. Were you to wish to guide people how they might judge others, you would create a very different document. Such a document may well be related the "broader issues" that RMS is apparently considering - it is certainly a "different" issue, though it is related.

> In other words: who do you think are the "I" and "You" in this instance?

"I" is always the individual who is creating the text or speech. It is also a person who approves a text or speech (as they are effectively repeating the text/speech and making it their own responsibility too).
So if I write a patch to the documentation, I would consider it appropriate to follow the guidelines.
If I were asked to review a patch, and found that the communication style of the patch (or the result of the patch) did not fit my understanding of the guidelines, I would think carefully about whether I should approve it.
I wouldn't say "you cannot say that, it contravenes the guidelines".
I might say "I don't want to approve this because I feel it speaks in a way that is discouraged by the guidelines".

Thanks for asking.