I, for one, am eager to see whether non-interactive video lessons, or interactive computer-based problem sets (both of which are available on Kahn's website) are successful for particular groups of students. We've heard lots of self-reported anecdotes of success from users of Kahn Academy, but there hasn't been any sort of rigorous, valid study of student outcomes, yet.
My suspicion (backed by my own unscientific, self-reported anecdotal experience as a high school math teacher in an urban school) is that videos and interactive problem sets work well, for highly self-motivated students, but that hard-to-reach students will remain hard-to-reach.
James Klock