My response (along with many others) was published in July's issue of Wired magazine (8.07). You can see it on page 70. It's short-- the entire text of my letter follows:
Joy says that the computational power will make smart robots possible by 2030 ("Why the Future Doesn't Need Us," Wired 8.04, page 244). But much cutting-edge AI research doesn't even tax the abilities of our current computers. The bottleneck is more in our understanding of intelligence at a process level than in the raw computational muscle.
Jim R. Davies
Graduate student
Intelligent Systems and Robotics
Georgia Institute of Technology