The initial criticism of Google's new knol initiative pretty much misses one of the biggest points. Everyone, even Ars Technica, makes the comparison to Wikipedia but ignores the biggest problem with Wikipedia that knol fixes: Wikipedia entries can be edited by anyone, not just the original author.
My contributions to Wikipedia have all be additions to existing entries and corrections to bad entries. That's fine, because I only post what I know about. But having looked through the edit history of some entries, I can see that lots of people who don't know what they are talking about add to and edit entries. That means that a previously-OK entry becomes worse until someone else fixes it (and that assumes that the edit in question doesn't become part of an "edit war").
On the other hand, the nascent knols will be from a single author. Google even said as much near the top of the announcement: "The key idea behind the knol project is to highlight authors." If you read a good knol and want to make it better, send suggestions to the author and hope that they incorporate them. "People will be able to submit comments, questions, edits, additional content, and so on. "
Knols will not be the death of Wikipedia. Wikipedia will thrive for readers who want the level of quality that it presents. I strongly suspect that, once there are about 10,000 good knols, the Wikipedia entries for those topics will either get dusty or degraded, but will still be useful to people who like the Wikipedia style. I also predict that knols will be quite frustrating to some readers who find good ones, suggest changes, and then see the changes not being made due to the author not having time (or disagreeing with the content of the changes). Those frustrated readers would do well to contribute to Wikipedia. Regardless, the two knowledge systems will be appreciated, often by the same readers.