Reply to comment

The disproof of Searle and the Kurtzweil Singularity hypothesis is really dead obvious: We have no acceptable definition of intelligence, so we cannot "create" it any more than we could create "smurtzwaggle zokelin" because the term is nonsensical. how could we set out to get to smurtzwaggle if we don't know what direction to begin or how to tell if we stray?

There is a related proof, and that is repeated underlined and highlinered in all AI experiments since SHRDLU: We ourselves do not understand even the very rudiments of how we ourselves learn, understand or use language. Anyone asserting the contrary is outright lying, filling you with fantasy Sci-Fi. We have studied the psychology of language for some 80 years now, and we have had a natural philosophy of the mind for many thousands of years, yet we are no closer to even the faintest understanding of how we do even the simplest of cognitive tasks. No idea whatsoever. We don't even have any faint understanding of where to look for the answer, all we have are volumes of rulebooks chronicalling medical anecdotes of where the language systems go wrong, and we cannot even simulate those with our programs.

Joseph Wiezenbaum, inventor of the famous prototypical chatbot psychoanalyst was amazed that even his own staff refused to believe that eliza was not understanding their problems; the desire to believe in AI is so strong, even in the face of huge evidence we cannot let go of the dream. Wiezenbaum also pointed out how even with our more stellar AI acheivements, in his day it was SHRDLU and in ours it is Deep Thought, you can blow the whole Mechanical Turk wide open by asking the genius machine only one question: "Will you sell me your bishop?" -- it is a question any two year old human will grasp and answer in an instant, yet will stymie the un-pre-pared machine.