Let me phrase this another way. If a predictive model is only predicting data about it's own model, it's going to be self proving. There is still a giant spac

Let me phrase this another way. If a predictive model is only predicting data about it's own model, it's going to be self proving. There is still a giant space of everything else. And the more ones thoughts become consumed with models of prediction, the easier it is to ignore the space of everything else. But I guess simple, boring language describing the phenomona isn't interesting enough. What is the difference between a cognitive model that 'predicts' it's own behavior, and a premonition, when none of the data has anything to do with anything besides thinking about thinking?