ANNABELL, the first neural network simulation (on a hacked GPU no less) to be “taught” how to “think”.  neural networks where mostly philosophy and math back in 1989 when i started.  who would have thought that gamers drove a hardware chip design (1500 cores on a off-the-shelf GPU card) that now gives enough wide-and-tall grid to do these fun things?  happy i still understand this technology now that the hardware can keep up to the dreaming.   some day, we will look back, and know ANNABELL was HAL’s mother.