I like the argument that Taleb makes in _Fooled by randomness_, which is essentially that most news has negative informative value because most of it is just noise. Only with time does the signal appear.