I have been meaning to post this. Lee Dirks, Director of Microsoft's Education & Scholarly Communication External Research Division, gave an incisive talk a couple of months ago at the OCLC/Frederick G. Kilgour Lecture in Information and Library Science. Dirks provided a pretty solid thinking through of the major paradigms we've got upon us in these times of big (and even chaotic) data. Technically and epistemologically his talk was pretty tight.
The video and the presentations (more than one) are available from the UNC SILS website: