ABSTRACT
We are now at the point where we can fully appreciate the inadequacies of conventional software engineering methodology when confronted with Al-ish problems—not to mention its inadequacies tout court. And those of us who have not given up on the possibility of practical Al software will probably concede the need for significant changes in the way that software systems are developed—perhaps changes so radical that, in effect, we are looking for a whole new way to go about building software systems, in a phrase, a new paradigm for engineering AI software.
