This short article in Business 2.0 references our experiment with blogs here in Utah and questions the effectiveness of using a search appliance like the one from Google to turn them into a useful knowledge base. While I'm inclined to agree in theory that categorization is probably better than raw search (that's why I use categories on my blog), I think that this argument is analogous to the Yahoo! vs. Excite (or Altavista or Infoseek) debates of a few years ago.
If you remember, Yahoo!'s claim to fame was human categorization of web sites into topics while the others relied on machine indexing. There's no doubt that the human categorization was useful, but it was also expensive. Eventually an indexing and ranking scheme came along (Google) that was good enough that the debate went away. Now most people use Google and its good enough.
That said, there are some categorization experiments for Radio that I'm anxious to try out.