I just skim read it. It looks like the brain stores a visual model and a semantic model. The first matches what is there for real time manipulation and the semantic model that gets stored for long term use like a cartoon version of what we see. Is this why cartoons work so well?