more AI -- lenat on common sense

if you were interested in the ai article posted by torless, (and you have a computer capable of playing windows media files) you should check out this lecture by doug lenat -- the founder of the cyc project.

it's been a while since i've seen the lecture, but i vaguely remember an extremely interesting section where lenat describes humans as being non-rational or trans-rational or something along those lines. he goes on to give examples of such non-rationality in human behavior.

i think this opens the door to a certain tension between human and machine intelligence. if humans (even intelligent and highly trained humans) are indeed non-rational, should we be trusting each other to make important decisions?

collaterally, an old and unfortunately no longer freely available nyt's article talks about research on how difficult it is for people to predict how they will feel in the future. they called this 'affective forecasting'. i have a copy of the article lying around, so if you would like to read it and don't have an nyt's subscription, drop me a line.