I just read this great example of how accurate Latent Semantic Analysis (LSA) is. LSA is a tough one, it doesn’t attempt to be human but with a large enough corpus can make solid similarity’s, solid enough that it would seem AI’ish. This eventually could be one of the products that has a hand in AI in the future. One might consider its maximal knowledge of the world to be analogous to a well-read Nun’s knowledge of sex, a level of knowledge often deemed a sufficient basis for advising the young. (This was written by Landauer, T. K., Foltz, P. W., & Laham, D. out of CU Boulder.)
