Please correct me if I'm wrong, but current applied NLP is building parse trees and applying probabilistic rules to classification problems. With those two yo

Please correct me if I'm wrong, but current applied NLP is building parse trees and applying probabilistic rules to classification problems. With those two you can indeed build some awesome tools but we are so, so far away from actual language that it hurts.