Well yes and no, then the question becomes one of what you are saying about what. This is a slippery slope because one can always argue that if the creator of something is still alive... you are studying people and I think we should explicitly avoid implication that studying people makes any given individual a research subject. There are a myriad of ways of studying people without fitting the description of a research subject in the u.s. guidelines (which is not as rich and plural as many humanist might wish, but ehh....). The situation in the u.s. as i've read it makes people a subject when you intervene, when you invade privacy, or when you might risk harm (harm in a way that is substantially more and different from the everyday.) I do agree that if you aren't sure whether you are subject to irb approval, you should get irb approval. But I don't think we should argue that studying blogs archived on the web and making inferences from their produced textual materials about bloggers in general is any different than when i study books and publications in order to make inferences about the operations of academia. Now, if i were to study a blog to understand one person's life that might be an issue, and it would depend signficantly on the methods you undertake as to where my opinion would lie.
If you are studying the site(s) as texts without consideration to the issues of the people who wrote the words, placed the pictures, etc....then I think you can use the text argument. HOWEVER, if you are using the words or pictures or whatever, on the screen to get at issues related to the content creators then you are studying people. PERIOD.