The section on not-arguing seems exactly in line with my thought on the matter. Well, I'd qualify "never tell others that they are wrong" a bit more - I'd say you should make sure it's factually true, and make sure that the person has practice in admitting mistakes in this area (think math students in math). But if those conditions hold, you can save a lot of time and condescension.
Anyhow, instead of making a post out of not-arguing like I'd half-planned, I may just fillet Carnegie.
In the very back of Kaj's excellent How to Run a Successful Less Wrong Meetup Group booklet, he has a recommended reading section, including the classic book How to Win Friends and Influence People.
It just so happens that not only have I read the book myself, but I have written up a concise summary of the core advice here. Kaj suggested that I post this on the discussion section because others might find it useful, so here you go!
I suspect that more people are willing to read a summary of a book from the 1930s than an actual book from the 1930s. What I will say about reading the long-form text is that it can be more useful for internalizing these concepts and giving examples of them. It is far too easy to abstractly know what you need to do, much harder to actually take action on those beliefs...