My background in AI

I was seen as a math/science type in high school in the 60s. During my first forays at college I took classes in Fortran and SNOBOL, and almost dove into computers as an easy career. But when I arrived for my first parttime job in the computer lab at Antioch, in 1973, I realized with perfect clarity that once I stepped into that privileged, artificial world, certain opportunities for self-discovery would be lost to me forever.

So I fled. But before I fled, I built my first FORTRAN model of human emotion-- a simulation of a college dorm, where students grew to like or dislike each other as the result of random encounters. We were looking to Jay Forrester's early economic simulations for inspiration, but what I took away from this experience was an appreciation of how ill-prepared we all were, to try and express human personality in terms of equations.

And for the next six years I stayed away from computers, and worked on self-discovery the way lots of people did in the 70s. And I was conscious all along that a computer model of personality was what I hoped to work out... but that the computer-part was trivial, or worse, distracting. The important part was the soul-searching! What finally roped me back into the world of computers was the appearance around 1979 of video arcade games like Asteroids and Missile Command. I had by this time worked out a primitive graphical story-notation system I called "Anti-Math" (2 gifs, 17k), and I thought if I could animate it with colored shapes on a video screen, it should be much more comprehensible to people.

So I taught myself Apple 6502 programming, and took a series of jobs doing arcade conversions for Apple, Atari, and Commodore micros. The animated Anti-Math got stuck when I raised my sights from 'dumb' story-recital, to wanting the program to generate new plots via planning-algorithms. But this new interest eventually led me to apply for a job at Northwestern, when I heard Roger Schank was moving there from Yale.

Working at ILS was a big break for me, and I learned a lot about AI programming, and a little more about what a hoax AI is, generally. I've written a memoir detailing a lot of this. And then in 1991 I stumbled on a new approach to representation, that I now call "fractal thicket indexing", and was shocked to find that the faculty and Schank were not just uninterested in the idea, but delivered an explicit ultimatum that I drop the subject forever, or resign! This is a story in itself, but ultimately I was fired, and a massive campaign of personal attacks was launched that included public accusations of stupidity, mendacity, and psychosis, and even threats of violence.

[Up to AI] [Up to JB] [Map] [No Next] [Robot Wisdom home page] (Feedback)