John Naughton provides the kind of really useful informal reportage that we need more of. I'll not try to summarise it, knowing that you will read it right away. One point that Hampton made that John questions was “The larger the haystack, the easier it is to find the needle.” I think that Peter Norvig's 2007 theorising from the data talk sheds light on this apparently counter-intuitive statement. If you've enough data then statistically based algorithms can home in the needles within the data, using other parts of the data to guide them.
Comments