Inefficiency of HTM topology from a ML standpoint

A post was split to a new topic: Merging topologies