And a logistic curve is even harder before it starts slowing. I think the various arguments for superhuman AI are far more significant than the projected curves.
The "logistic growth curve" charge, as voiced by John Wilkins:
First of all it occurs to me that people who expect the Singularity to occur simply do not get the logistic growth curve.
Here's a great article by Paul Allen about why the singularity won't happen anytime soon. Basically a lot of the things we do are just not amenable to awesome looking exponential graphs.