Like you can be a very friendly, chill, and fun person, but if professionally you keep overpromising, not own to your mistakes, pretend you know something when you don't (which could make you look cool outside of work but a pain at work), and other negative traits discussed in the paper, then I would avoid interact because you're not work-pleasant.
This describes exactly the type of people the corporate IT department at my company. Really pleasant people until you actually need something.
Career wise these people can go very far though.
To me it's an indicator that the "simpler" is that management (and possibly more, given concomitant hiring practices) has only one way they know how to interact with people: after-class mode from college. In Social FizzBuzz it scores a "Fizz."
I'm not going to disagree with you re: beer buddies. Really the way I read this is not as a requirement that employees participate in after-work events or in mandatory work friendships, but more of a misguided heuristic for something simpler. The general question is whether the average interaction with you is pleasant or not. If your job requires interacting with others (as it often does in software teams) and you're generally unpleasant or difficult to interact with, then people might avoid interacting with you, and that's a hit on everyone's ability to do their job. On the flipside, someone who's easy to interact with can be a great resource to everyone and also learn from all the people they interact with.