On a side note, Geoffrey Hinton just told in an interview that he used to think that it's 20-50 years away (as I said I've never heard a specific number from anyone before this) and now he thinks it's less than 20.
Now add to this that people working in the field tended to systematically underestimate the progress. My gut feeling is that if Hinton says less than 20 years then it can easily turn out to be 10 or less. (Also, I very much respect him for coming up with an actual number.) I think AI researchers underestimate the progress partly because they don't want to look stupid by making over-optimistic predictions. However, in this case, I'd argue that the optimistic prediction is saying that it takes longer. Because it will definitely stir up the whole society and economy at the very minimum.