The best thing I have ever read is von Neumann's ridiculous ideas and predictions about the weather.

It is the ultimate example of always having to be on guard against argumentum ad verecundiam.