The distinction is between an open knowledge network, a private knowledge network, and a mixed open/private knowledge network. I'm promoting the idea of a mixed open/private knowledge network rather than entirely open or entirely private.
So by private I mean access controlled. This would mean algorithms which are run in such a way that no one can reverse engineer or know the algorithm. This would mean code which can run without the users knowing what the source code is (closed and private). There are pros and cons to this though because while the benefits are clear (intellectual property being an example) there are also risks (malicious code). So there has to be a way for the community to balance out the risk vs reward and for each individual to determine the level of risk they are willing to accept.
For more check out the links to the Enigma forums where the philosophical issues are being discussed.