I like the thought provoking concepts this proposes, though it feels very unrefined. It seems to propose a black box where you have a vast collection of arbitr

I like the thought provoking concepts this proposes, though it feels very unrefined. It seems to propose a black box where you have a vast collection of arbitrary data, you give examples of data you want and don't want, and you get arbitrary outputs. I wonder how something in this day and age can use labeled examples and machine learning to find more examples to classify portions of data for use in a search engine. Certainly something that Google has probably experimented with or actually uses.