samwise.......

you're looking at the estimated/predicted temperature increases vs CO2 increases, including feedbacks.

The graph that John presented was the temperature increases (calculated) vs CO2 increases, without feedbacks.
The italics are the critical difference here.

It's fairly well documented that from first principles, excluding any changes in convection, a doubling of CO2 will yield a temperature increase of 1.1-1.5 degrees C. You have to include some silly positive feedbacks, and minimize any negative feedbacks to get the 6oC increases predicted by some GCMs.