In today’s NYT, computer pioneer Jaron Lanier presents lucid and searching counter-argument to the artificial intelligence regimen that has dominated too many corners of computer science for too long.
He dismisses the notion that eventually the combined intelligence of all processing resources online will evolve into a super-intelligence that will then recode humans into a more serviceable form (the Kurzweil/Singularity vision.)
I concur with Lanier that we need to focus on extending the boundary about how to capitalize on the immense potential of human beings as social entities and design systems that augment and extend these capacities. He also warns that the comp sci community has imbued new developments with a religiosity that will put their developments at odds with conventional religious disciplines. On that point, I understand the concern, but a full discussion of these issue is, as they say, beyond my pay grade.
Bottom line: just read Winograd & Flores. That book (from 1986) frames this discussion impeccably. Augmenting human intelligence stems from augmented human communicative capacity.