During that meeting and ones that followed, the full scope of the project and why the military were funding it became clear.
At first, it seemed that the military wanted a system for making predictions about foreign political climates based on social factors. An enormously complex task, but extremely rewarding intellectually to work on. The immediate applications of this were obvious, even to my somewhat naive mind. The political climate of a country affected many things, from their foreign policies to their economy. As part of a greater system such predictions could provide valuable insight into the workings of a country. Of course, by making simulations and predictions, one could experiment with deliberately influencing a country in whatever way one chose, playing with the different factors until a desired result was achieved.
I spent several months evaluating the different factors involved, trimming the ones that were for the most part inconsequential and working with other teams on developing ways to actually record the data needed for the other factors. It's all about information and it's hard to record things that are inherently unquantifiable, but that wasn't my job. I was just part of a greater whole, a single cog in a huge, whirring machine, but don't think that I tell you this as a plea for absolution. I knew what I was a part of and if I had chosen to think about it, I could have seen where the research was headed, but I didn't want to think about the applications. I was hungry for the challenges, the chance to flex my intellectual muscles and put my lifes study to real, practical applications. I was an idealist, I never stopped to consider the moral implications, taking it for granted that the knowledge gained from our work would be put to good use, would benefit mankind. In their own way, I'm sure the military thought the same thing.
The project continued smoothly for the next few months before we hit a snag. The model the system was building was changing to fast for it to cope with and no matter how much horsepower the techs threw at the thing, it would just consume what was given it and then grind to a halt once more. We dared not reinitialise the model. By this point, The Oracle (we had taken to calling it that ever since some tech had told a joke about the Oracle at Delphi I didn't understand) had been designed to take advantage of dynamic neural networks, it was in a very real sense a very large, virtual brain and there wasn't an easy way to just take a snapshot of it, turn it off and turn it on again. At least, thats what the techs said, it wasn't my field.
Eventually, I came up with the idea of modelling it more like a human brain, with long and short term memories, a system for relegating obsolete and often unchanging facts to lower level functions. I also incorporated some of my theories about the collective unconscious. After a period of slowly introducing these concepts to the techs, they ran some simulations and were excited by the results. We got approval from up top to push on forwards and so we did.
It was at this point the project began to take on a life of it's own.