I'm about to start a PhD in computer science. One might also think that CS research is something where you just live anywhere you want and solve interesting theoretical puzzles all day, but it's surprising to see how much communication is important (both written and oral). You need to tell people about your work and understand what other people do, but most importantly it's crucially important to think about hard problems together with other people.
It's true that you can work on your own, but (at least from my perspective) it's incredibly more productive to think about hard problems with someone. It forces you to spell your ideas fully in order to communicate them to the other person, which often deepens your understanding of the ideas (or makes errors obvious -- think of rubber duck debugging): it's true that you can do it through pretty much any medium, but face-to-face is still the best choice by a large margin. Besides, face-to-face has the added benefit that when you have someone else in front of you, there is a sense of urgency that prevents you from turning your brain off, whereas if you're thinking by yourself it's much easier to lose focus.
For this reason, even though it seems that you could do CS research your own, a vast, vast majority of papers (at least in my field) are collaborations, and 1-author papers are incredibly rare. Besides, though researchers are fond users of email and Skype and other such tools, they also go through the time and expense of moving to remote countries just to be able to collaborate face-to-face with the right people.