Thomas K. Landauer: “if you want the machine to agree with the humans, why not just ask the humans? Everyone would be happier.” Landauer's argument is compelling because he seems generally skeptical about machines doing work humans can do more effectively, if not more efficiently, a skepticism I share. (He also has a financial interest in arguing the opposite of what he's arguing, making his actual argument even further compelling.) I'm trying to avoid what seems to be an likely trend towards computers aggregating and filtering out weblogs based on relationships and other criteria, where having a human do it at least gives one a window into that human's personality. Eventually artificial intelligence will be indisguishable from human intelligence, but I'll always want nothin' but the real thing, baby.
Why Not Just Ask The Humans?
