highwaystar101 on 04 February 2010
| Rainbird said: Wow, that's pretty awesome. I wish they had asked some more questions for more accuracy though, but it's still pretty awesome  |
In the paper they published about this, they justified only asking six questions as an more than sufficiently accurate result. Essentially what they had to do was pass an augmented Turing test, and six complex questions is sufficient enough to prove that you're communicating with a human.
In all fairness, if they were wrong about their hypothesis the results for each question should have been null (neither yes or no).
