baez@x.ucr.edu (john baez) (11/11/89)
In article <MZHXDmS00W0JI=l1cj@andrew.cmu.edu> eh1s+@andrew.cmu.edu (Edwin Huang) writes: > I'll try to summerize the discussion. Aparently there is an ainu on >comp.object. Ainu by the way stands for Artificially Intellegent Network >User. The AI supposedly was developed by MIT and posts with the name of >Ed Berard. Is this real????? From a friend who shall remain anonymous: I doubt that Berard is purely AI. Things which are conceivably plausible to me are: - it is AI but guided at a high level by humans, AND - it scans a huge database of existing news articles (or some other source of essay-like data) for new "ways of saying things". Of course, I think it is most likely that it is purely human. I don't notice much non-human about what it writes, and I notice a lot of human things that are hard to explain any other way. For instance, in the first article (with 100-line gaps), it uses several "semi-poetic" ideas, like "since I'm a scientist here's how I planned to attack this", which to me has a Lem-like humor in applying a procedure in a straightforward and even reasonable way to a situation that most people would not think of using it for, since it does not seem straightforward enough to be procedurally handled. It is an intriguing thought that an AI system might think of something like this, not being blocked by the human knowledge there was something a bit silly about it! But to be convinced of this, I would of course have to have the code to play with, or an explanation of how what it says could be generated. I don't consider the object-oriented programming article rambling at all! I think it is highly and not very creatively organized, well organized, and I'm not sure if it's organization is sensible enough to require human input.