In most White Houses, policy and action flow down, with staff trying to implement what the president wants—or, at the very least, what the chief of staff says the president wants. In the Trump White House, policy making, from the very first instance of Bannon’s immigration EO, flowed up. It was a process of suggesting, in throw-it-against-the-wall style, what the president might want, and hoping he might then think that he had thought of this himself (a result that was often helped along with the suggestion that he had in fact already had the thought). Trump, observed Walsh, had a set of beliefs and impulses, much of them on his mind for many years, some of them fairly contradictory, and little of them fitting legislative or political conventions or form. Hence, she and everyone else was translating a set of desires and urges into a program, a process that required a lot of guess work. It was, said Walsh, “like trying to figure out what a child wants.” But making suggestions was deeply complicated. Here was, arguably, the central issue of the Trump presidency, informing every aspect of Trumpian policy and leadership: he didn’t process information in any conventional sense—or, in a way, he didn’t process it at all. Trump didn’t read. He didn’t really even skim. If it was print, 1t might as well not exist. Some believed that for all practical purposes he was no more than semiliterate. (There was some argument about this, because he could read headlines and articles about himself, or at least headlines on articles about himself, and the gossip squibs on the New York Post’s Page Six.) Some thought him dyslexic; certainly his comprehension was limited. Others concluded that he didn’t read because he just didn’t have to, and that in fact this was one of his key attributes as a populist. He was postliterate—total television. But not only didn’t he read, he didn’t listen. He preferred to be the person talking. And he trusted his own expertise—no matter how paltry or irr