First you have to define what you mean by "genius", preferably in some way that is at least generally consistent with the commonly understood meanings. Then you

First you have to define what you mean by "genius", preferably in some way that is at least generally consistent with the commonly understood meanings. Then you have to assemble an unbiased set (very difficult) of people that fit your definition. Then you have to see what styles of thought that you didn't explicitly or implicitly filter for by your definition are also largely (and disproportionately) present in your set of geniuses. Then after you have found you styles of thought dominant among geniuses, you have to see if they actually work as predictors. I don't feel that the author of this article did these things.