I'd originally planned a post on The Golden Compass, which I'd started reading months ago. I didn't want to write about how horrible I thought it was until I'd at least finished the third book, but my travel schedule interrupted my reading after the second and I couldn't bring myself to read it, so I kept putting it off. Fortunately, that was covered by others and it doesn't look like it's doing well.

I'm glad, in some respects, that I didn't. I don't post often, but when I write about movies, I want to write more about movies that Conservatives will like. We don't give enough credit to the (few) movies that we have good reason to appreciate. It's too easy to criticize the vast array of movies we don't like. So, instead of Compass, I went to see Atonement, which stars Keira Knightley and was directed by Joe Wright, the same team that made Pride and Prejudice a few years ago, garnering Knightley an Oscar nomination and three ticket fares from me, the only time I've seen a movie multiple times in the theatres since I was little (all viewings alone, I'll add. Go ahead; you won't be the first to question my masculinity). I'd seen no promotion for it; it was only when I was looking up movie times for another film on Fandango that I saw an ad for it in the banner.

I loved the movie. It feels like a revolutionary concept these days, but honor matters. The movie is about how courage is not just for heroes and warriors, but that cowardice can have terrible consequences between people in all of our lives as well. I'm not saying this is not a Liberal idea, but it is definitely a Conservative one; if Liberals like the movie, great. If they like it for the same reasons, even better.

It also struck a rare balance by depicting war as something dangerous and hard, but not in a way that condemns it solely on the basis of being dangerous and hard. Like Thomas Sowell says, whether or not one supports the current war, nobody is actually "pro-war". War should be a last resort, because it is a horrible thing. Some of us simply think it's not an unthinkable option and that the current situation demanded it. I appreciated what Mr. Wright did here.

It's got some dirty words, not used in dialogue casually, but used as something dirty and vulgar, which is an improvement over, say, Romancing the Stone, for instance. Also, some full-frontal nudity. Well, that may be a slight exaggeration, but a throughly wet slip is not exactly opaque (I think I just sold the film to every 14-year-old boy in the nation) (except they don't read this blog). Also, I wouldn't see it if you have loved ones at war; wait to see it with them when they return.

It's too bad that it doesn't seem to be doing well in the theatres; I wish I'd written this on Friday, except that I don't think it would have done much good. Well, it did get me to finally open and watch my copy of Pirates of the Carribean: At World's End, which I'd bought on Tuesday and haven't had time to watch, and to (finally, again) purchase a copy of Bend It Like Beckham, which I saw in theatres back in 2002 and also loved.

Of course, I already own a copy of Pride and Predjudice.

So, in the early days of computing, Alan Turing, one of the most influential Computer Scientists of early computers, proposed the "Turing Test" for AI systems. Basically, it said that, if you are interacting with it and you can't tell whether you're interacting with a computer or a human, that's true "Artificial Intelligence". The problem was, it was beaten pretty quickly by a program called "Eliza", named after Eliza Dolittle from My Fair Lady and Pygmalion because they taught it to talk.

Basically, Eliza assumed the limited social context of a Rogerian psychologist and attempted to imitate the psychologist. You know, you say "I hate books" and they/it responds "Why do you hate books?". Well, if you look, it didn't even change the input very much, it just kicked what you said back to you (think the Lucy Liubot episode of Futurama (3.15)). Eliza licked the Turing Test hollow, but it was the stupidest program you can imagine. A few hundred lines of code and no actual knowledge of psychology and barely any of grammar. Eliza is a very famous program, not because it's AI, but because it showed the Turing Test wasn't much good, not because computers are smart, but because humans can be pretty dumb. It turns out that, if the social forms are rigid enough, it's tough to tell the difference between a smart computer and a dull human. In fact, Eliza was so famous, when I first learned about IMbots (like the WSJ or Moviefone bots), my plan to build one was to filch somebody's Eliza engine and hook it up to an AIM account.

Well, somebody beat me to the punch (HT Instapundit), except they didn't mimic a Rogerian scientist; they mimicked members of a dating chat room and used it to steal personal information. The AI researcher in me (a former life, as computers go; 5 or 6 years at least) is curious to know if the social rules of flirting in a dating chat room (which, given the limitations of text, I assume are pretty rigid) are more or less rigid than a Rogerian psychologist. Did they just change the rules in the ruleset, or did they have to do any actual work?

Then I realized that I don't really care enough to investigate. It's funny enough as it is.