Computers versus common sense (abstract)
Douglas B. Lenat · 1993
It has been eight years since we began the Cyc project—a massive decade-long enterprise to codify, once and for all, the millions of general rules of thumb about the world (that are generally held by adult, rational members of their culture.) The motivation for this heroic effort is to produce a new “layer” between the Operating System and Application programs —a layer containing general knowledge and inferencing abilities, which could facilitate knowledge-sharing across applications, increase the robustness of an individual application, enable a new generation of natural language understanding programs, user modelling programs, machine learning programs, etc. Due to a large number of errors canceling each other out, we are still essentially on schedule. A report of some of the lessons learned so far, and accomplishments to date, should prove relevant and interesting. Moreover, many applications of Cyc are underway, and we will briefly describe a few of these. But even more generally, there is a lesson to be learned here for all of us: not to be blinded by the “physics envy”, by the word “science” in “computer science;” the lesson that sometimes a large real-world problem may admit an engineering solution but no small, elegant mathematical solution.