The second day started with a keynote from Peter Norvig which was quite entertaining, too. Below you can see the essence of his talk and what he thinks what are the prerequisites of Artificial General Intelligence (AGI). In short: Probabilistic First-Order Logic over lots of data. First it sounds like that maybe only Google has access to "this lots of data", but now they share it with us. You can get approx. one trillion word-tokens (n-grams, from unigrams to 5-grams) of text from publicly available web-pages in 24GB gzip'ed text files over the Linguistic Data Consortium.
In the Q&A session someone asked Peter if he has maybe observed "emergent intelligent behavior" in the large server networks of Google. I am not sure, if this guy was just joking...
The other interesting part was again the american pragmatism: American Venture Capitalists are already interested in the possible business created by the Singularity. In germany, this is unthinkable. I guess most german VCs don't even know what the Singularity is. And if they knew, they would think about it as science fiction... or scientists gone mad.
So I agree with Rodney Brooks: The Singularity will not be a moment in time. It is a period. And it is not only about intelligent machines, but also about human transformation. Maybe the last people who will be transformed are the german VCs or our government (we shortened the distance to russia and singapore).