First, my sentence was carefully constructed to say that his essays "do not comprise a collection of disparate examples surprisingly linked through a common thr

First, my sentence was carefully constructed to say that his essays "do not comprise a collection of disparate examples surprisingly linked through a common thread." I scanned through a couple essays to confirm my impression, which is that his essays are mostly composed of his personal experience plus directly-applicable examples which are used to illustrate a point. Secondly, I said ideally statistically valid. Good essays can be written without the use of statistically significant data properly interpreted, but the conclusions can only be as powerful as the evidence used to support them, as always. If Paul Graham's essays had included, for example, statistical data on the performance of startups that he has funded broken down by category I would consider them more highly informational. What's going on here is that I think the level of insight and analysis that many people think is laudable in a written work, I consider shameful, and highly prone to bias and error. I never said I don't agree with the thesis of this piece (I have absolutely no dog in that fight). Perhaps you think I am the same person that you were originally responding to?