Brainhat is a pet project of mine. I began coding in 1996, when I was still running Atlantic Computing. I would find time on weekends, during flights, vacations. We exhibited at SpeechTek for a number of years. Brainhat became an organization in its own right in 2001. I financed it initially with money derived from the sale of Atlantic to Integralis. Dan Collins, Intregralis' ex-COO also devoted some funds, and we were able to continue officially for a while. Unfortunately, the dot-com bubble burst and I had to go find real work again. So what have we accomplished? We have developed brand recognition and proprietary technology within the industry. And we advanced the state of the art; Brainhat truely represents knowledge, runs inferences and generates output using semantic structures. We can use English as an associative programming language. And we have interfaced to everything that moves. Brainhat is real Artificial Intelligence. We've done amazing things.

The source code is available. Brainhat is probably best suited as a next generation, goal-oriented command and control platform for robotics and service tasks. It would be good as the basis for a conscious house or inferential business security platform. Imagine computing with verb tense, association and imperfect data and hypotheses—the way people think. Now imagine an army of little brainhats chattering and sharing what they learned—the results of their inferences—in a parallel broadcast domain. The Gestalt of it all!

At the same time, I am working to make this set of programs easier to integrate into real products, more efficient and language agnostic. An occasional intrepid student or researcher will take up the code. I'd be nervous about making it too approachable, as I don't have time for the technical support burdern. But, it could be easier... I have to re-address the speech interfaces too; they're suffering from entropy. My ultimate goal: a million new consciousnesses.

write: dowd@atlantic.com

Kevin Dowd