It is something that drove me crazy when I wanted to follow political events from multiple perspectives (I used to be a political scientist, so this was an important part of work). Google was both a great friend and worst enemy: I don’t want to follow obvious and predictable news sources, but gauge as broad a perspective as possible — maximize variance, without regard for the mean, so to speak. But Google was focused too much on the mean and was trying to feed my search results with the news stories that its algorithm thought I wanted to see. Predictive analytics are nice, but, sometimes, noise is the important part of the data, not something to be brushed aside as nuisance — which is an attitude that I’d run into too often. The important thing is to recognize when the noise is just noise and when it is actually an important part of information, I imagine.