Someone needs to keep track of these types of predictions and track how often the author is wrong. It is costless to make big predictions, because rarely do people come back and say "You were wrong." It would be nice if I could search for this author's forecasting performance.