For example, it doesn't seem to have indexed much from ICML since 2010 (though, curiously, it indexes exactly 3 papers from ICML 2011): http://academic.research.microsoft.com/Conference/35/icml-in...
Similar story for AAAI: http://academic.research.microsoft.com/Conference/251/aaai-n...
One thing that's neat about GS is that it has a negative lag instead: it's already indexing papers from a bunch of conferences that haven't happened yet, and whose proceedings aren't officially out. It finds the versions that authors post on their own webpages, or in their institutional repositories, almost as soon as they're posted. For example, you can already read a good portion of the CHI 2013 proceedings if you search on the ISBN: http://scholar.google.com/scholar?hl=en&q=%22978-1-4503-...