if we’re talking about the “semantic web” as basically just another “non-intelligent” tool for intelligent people, then tools such as mindmanager (as any other content editing/linking tool) might indeed be helpful for building it (why shouldn’t it be? ;)
my take is (as stated earlier), that this kind of “semantic web” is possibly fine for the next say ten years or so. it’s a rather pragmatic view of the “semantic web” but as a matter of fact this view suffers from an old-fashioned approach which will never allow the “semantic web” reach a higher level than that of just being a tool. (there are many problems with the current approach such as the frame-of-reference, categorization, generalization, adaptation, symbol grounding problems, lacking emergence etc.)
if however we are interested in making things intelligent, a new approach is needed. this approach should try applying some of the principles of “new ai” (brooks, pfeifer et al.) in “virtual worlds” (true “new ai” advocates may shudder when reading this ;). i have some ideas how to start with this.
a good starting point for anybody interested in this field is “understanding intelligence“.