If the humans are actually trying to prove their human-ness (like they get a prize for convincing others that they are human), then I cannot believe how any bot can get >1%. I see these articles a lot over the last few years, and every time the bots fail to simple contextual questions and general conversation flow.

That the winning bot got 30% tells me that the humans in this experiment were not trying their hardest to appear human, and I don't see the point of a Turing test where this is the case.

TFA links to the actual list that Oracle claims, why isn't this in the summary so that people can judge for themselves:
http://constellationrg.com/blog/2012/06/oracles-list-100-plus-cloud-applications

I don't see why this is news though. Marketing department chooses a convenient level of granularity to get a shiny number. Isn't that like, all marketing?