Your last point is the most persuasive. It's possible to make all sorts of predictions using models, etc., all valuable. But there is too much that we just don't know. Mucho shit will happen. Probably already is happening.