[2b2k] Correlation's diminishing returns

Jonah Lehrer has a terrific article at Wired about the limitations of the causality model as data scales up and becomes more complex. (I'm over-simplifying to the point of inaccuracy.)