While we are at it, there is also the 2009 paper: The Unreasonable Effectiveness of Data by Peter Norvig.