Now here's an intriguing game, called John Conway's Game of life:

Basically, you feed it a certain pattern, and set it into motion to see how it plays out. You usually end up with a seemingly stable end result, with a number of stable configuratios that last, and a number of flip-flop situations. But the slightest new input might thoroughly disturb the precarious balance, and result in a storm of epic proportions, where large areas are irrevocably changed before settling in a steady state again. Just observe how lack of inputs always seems to result in an eventual static plus dynamic part that remains until the next input.

I am currently pondering the application of such an array as a neural network, but then we'll have to change the rules of the game a bit. Just imagine a cubical space, of cross-connected cubes, where every single cube has six sides, and where the color of the sides is dictated by the values of their corners. so each corner would have a byte value, and together they form a standard 32-bit color code, as used in today's computers.

Imagine how a 3 by 3 layout would look much like a rubik's cube, but with it's sides having colors that are never the same on one side! Inside, the cube would exhibit waves of colors travelling through the medium much like the grid patterns of John Conway's game of life travel through their 2D space. Do you know Conway's game also has so-called producers and absorbers, that either fire projectiles into the grid, or absorb them? I figure these structures will exist in the 3D version as well, along with other more exotic waveforms.

But now just to figure out a proper transferance function for inside the cell: it has eight corners, and how does it transfer information that is incoming into useful next values? That will be my focus for today....