Ask HN: Any one working on the "Hierarchical Temporal Memory" approach?

This is the HTM theory proposed by Jeff Hawkins of Numenta. Can anyone share their thoughts on Sparse Distributed Representations approach, vis-a-vis the approach-du-jour deep neural networks. Have you had any success?