Well, I was going to post one from Ramona Ausubel’s “Atria” from the 4 April 2011 New Yorker, but every paragraph was very good. Every single one. Damn, it’s such a good story. So instead, we have to go with this paragraph in the Adam Gopnik essay, “Get Smart”, on recent books looking at computers and A.I., which paragraph is also very good:
[The author of one book]’s central point is that the Turing bots that work best, whether produced by a computer or shaped by a mind, have to be, or fake being, dynamic. The best test of their humanness is not how smartly they offer answers but how quickly they interrupt, get distracted, compress information into slang codes, rely on “uh”s and “ah”s. Intelligence is an affect engaged in an activity. It flits between the empty spaces as much as it takes place in the exchanges. If a teen-age boy says to a teen-age girl, “I was, like, wondering, if, like, you’d like to, like, go to that, uh, thing at Jacob’s?” and she says, “Uh, well…” it’s bad news. But if she says, “Well, um…” it’s promising, and if she says, “Yeah, like, funny, because um…” it’s the best news of all. Prefixes and tics and characteristic mannerisms are richly coded with information. The two best Presidential communicators of recent decades had distinctive vocal prefixes that did a lot of their talking for them: Ronald Regan’s “Well…” meant “Despite your attempt to antagonize me, I’m still going to appeal to plain old placid common sense,” while Obama’s “Look…” means “Forgive me if I sound impatient, but if you actually examine the facts in the case you’ll see I’m right.” One marker assures Capraesque cheer the other, Spockian certitude. And it’s hard to make either understandable to a machine.