Re: Sarcasm Detector Wanted
There's a famous notion about how computers can't play the traditional Japanese game of Go. Supposedly, it's not because the game is too complicated, but because it's too simple. You can place a piece anywhere on the board, at any time. Apparently, that's too little complexity for any computer to derive a winning strategy based on. Now, I have no idea if that anecdote is true or not (don't there exist Go video games?), but I think it's very relevant to the discussion at hand. At a certain point, do we run up against a wall of what computers simply can't do? Or are there endless possibilities, some of which we just haven't unlocked yet?
Maybe the secret service needs to go back a couple of steps and get a human sarcasm detector first, though. Sitting out in front of your grandfather's house for a letter he sent years ago, especially considering his age at the time, was more or less a waste of taxpayer dollars. We ought to work on automated criminal detection in the sense that the volume of tweet (etc.) are too much for humans to read... but we still ought to have a system of humans with common sense to back it up. The debacle with the NYPD twitter campaign you wrote about back on E2 is a great example of this problem and how it goes both ways. Anyway, here's to four more years of Geekend!