I think it's kind of fascinating that the description of the culture of the "old, dank pub" could equally apply to Emacs and other long-lived free software projects.
One reason might be that it's a style preferred by the creator.
Yes, it could be a purely textual conversation between two people but, for example, the textual expression of the second-to-last panel might simply be:
Alice: ...
Bob: ...
Certainly there is a point that a person does not have to see value in anything other than the words but I don't get why it should be so perplexing for a person to see value in the illustrations.
(I say perplexing because of the phrasing, "Why even bother ... ?".)
Because it's a comic. Comics involve both illustrations of characters and those characters speaking, so he draws characters instead of just text.
If you follow SMBC this is a pretty normal setup for him, and many of the comics are the same panels with different text. This happens constantly, with the most notable example being two girls hiking up a snowy mountain. He's not quite as egregious about it as Dinosaur Comics, where literally every strip is the same with different text.
This comic was posted here because the text was particularly interesting this time. And it's not just talking heads, he also gives the illustration of the intersection as another visual metaphor.
