I didn't like one aspect of the book, when he was trying to show that what simulated people feel doesn't depend on the way simulation runs, he was mainly focusing on computing later timeframes of the simulation before earlier ones, which is not possible.