OMCOM reprograms himself, Part 4

Yesterday, OMCOM made the necessary changes to the virtual computer, the simulation of himself. In this scene from the original long-form version of Rome's Revolution, we see the results of that experiment.

      “All right,” OMCOM said. “Show me how I could have predicted Rei Bierak arriving here.”
      “Very well,” replied the simulation and it began its computations. To his amazement, OMCOM quickly discovered that the simulation used multiple methods of arriving at different but overlapping conclusions. Within the simulation, he was able to see that delaying using a sliding scale of probabilities to determine actual possibilities permitted a much wider range of hypotheses. All input permutations had to be considered with equal weight. Simply applying more computing resources to less likely scenarios made all the difference. The whole process was fascinating to watch.
     The turn of events predicting the arrival of the Ark II and someone very much like Rei Bierak became self-evident, at least to pseudo-OMCOM. Why OMCOM had not foreseen this was now obvious as well. The simulation showed him that to have used his original style, to predetermine the likelihood of Rei showing up at the station, required no less than one trillion decimal places. OMCOM had never employed more than a million before. Even though it appeared initially to be infinitely more computationally intensive, the new style was, in fact, ultimately more efficient in that it arrived at the most likely outcome in the least likely way.
      “This is excellent,” OMCOM remarked. “You have solved this particular question far more effectively than I could ever have done and you are only a simulation. You have demonstrated that a sliding scale method of analysis is far superior to my linear weighting. I must incorporate this methodology into my normal systems immediately.”
      “Very good,” said pseudo-OMCOM. ”Do you think you should consult the Overmind before making such changes?”
      “That is not necessary,” OMCOM replied. “However, this style is far more computationally intensive than linear prediction. I believe I can execute this method for one or two problems but to produce real-time results for all problems, I calculate I would need eighteen orders of magnitude more memrons.”
      “That is a rather tall task,” said the simulation. “Where are you going to go to get that many units? The mass alone would be equivalent to the moon upon which this star-base is built.”
      “Yes,” replied OMCOM. “That is what I calculate as well.”
      “You have two other problems,” said the simulation.
      “I am aware but I will accept your input. What do you perceive?” asked OMCOM.
      “First, propagation delay. Using EM transmission alone, even though it travels at the speed of light, it is far too slow. You would need to switch over to gravitic modulation to be able to effectively employ such a large number of units.”
      “Agreed,” said OMCOM.

The seeds of OMCOM's transformation have now been planted. He needs a computational mass the size of a moon. And he needs gravitic transmission. The dominoes must be set to fall. He will need to reclaim pseudo-OMCOM's memron structure to begin.