Richard Hsu has a post that talks about the differences between T-SQL's “Not In” and “Not Exists” functionality, stating that “Not Exists” can be faster. SQL Server Books OnLine talks about the two in generic syntax details, as well as some of the symantec differences, but doesn't really get into the gory details of which is optimimal in what situation.
I'd like to learn more about the differences between the two. If anyone out there can point to some good resources that talk about the differences in performance, etc, please post a comment here.