Intriguing. It's hard to imagine any creative work of art or engineering that couldn't benefit from well-curated annotation. Browsing around Rap Genius, I can't help but wish for a digital copy of Ulysses with this kind of interface to crowd-sourced commentary.
But outside of big traffic draws like song lyrics, I'm not sure what the monetization potential is.
There are incredible online annotations for novels by Thomas Pynchon[1] and David Foster Wallace's Infinite Jest[2], which are what I thought of first when they talked about expanding into annotating literature. The monetization potential in that area in particular is huge, looking at Cliffs Notes and the like. It's interesting that Andreessen Horowitz is investing in a company that seems to be solving the community problem rather than the technical problem of mass annotation. It seems like the technical challenges were the main hold-up back in the early 90s, but Wikipedia has shown the power and importance of community building when you're trying to create a universal reference.
Song lyric translations are huge outside of english speaking countries. U.S. rock/pop is played all over the world and most people can sing some words of a well-known english song. The potential to reach an international audience is huge, given the current deplorable state of other lyric sites. Getting a literal translation doesn't really help with song lyrics given that it tends towards poetry.
Good annotation for potentially ambiguous text documents is definitely monetizable in other areas. Look at SparkNotes and CliffsNotes in education. Being able to skim verbose or confusing sections of, say a physics textbook, with simple explanations from other readers with comparable backgrounds to me is incredibly helpful.
But outside of big traffic draws like song lyrics, I'm not sure what the monetization potential is.