It sounds cool, but they don't show any performance benchmarks. (In an interesting demonstration of both how fast this field is moving and how it's limited in part by computational constraints, they do say they're working on one... but it's only trained partway as they were writing this up!)

This is similar to how the cortex routes information through the 6 cortical layers - a cortical column. (http://en.wikipedia.org/wiki/Cortical_column)