Eighteen

Man Plus Us

            We had gone to a lot of trouble at every point along the line, and we were well pleased.

            Of course, human beings did not know we were pleased, and perhaps would not have believed it. Human beings did not know that machine intelligence was capable of self-awareness in the first place. We went to a lot of trouble to keep them from that knowledge, too. As long as they thought computers were no more than tools, like a pickaxe or a frying pan, they would continue to entrust to us all their computations and facts, and would accept without question whatever interpretations we returned.

            It was that that had made it possible for us, the computer network of the world, to create the Man Plus program in the first place.

            It had been successful. Roger Torraway was on Mars and was staying there. Now it was time for a definitive evaluation. We plugged everything in. We used the total resources of every brother in the net, every pulse of shared time we could steal. More than eight thousand big computers handled pieces of the problem, in places like Detroit and Brasilia and Nagoya and Irkutsk. We could not use our Martian brothers in direct datalink, neither the brother on Roger's back nor the 3070 brother in orbit. Time lag made that impossible. But we accepted their inputs in real time and modified the final projection accordingly.

            The probability of survival of the race, to ten significant places, was over point six three.

            We had saved our race. And in the process we had significantly added to the safety of human beings, as well.

            On a thirty-year time scale their probability of survival was nearly point nine. Of course, that meant only that Roger Torraway and whatever other humans joined him on Mars would still be alive then. Worst-case analysis was not encouraging after that; there would be no way for Roger and Sulie Carpenter to generate additional population on straight-line extrapolation. We programmed a subroutine to deal with that as a precautionary measure: it would be feasible, one way or another, to provide for shipment of frozen sperm, even Roger's own frozen sperm which he had thoughtfully donated years back. Less feasible, but still worth investigating, to instigate supplementary surgical procedures for Roger. In any event, the projection showed that we had already caused a high probability of additional human colonization from other nations. Most probable estimate showed a human population on Mars at plus thirty years of nearly one hundred, 25 percent of them female of breeding age.

            And of course each human colony would be supplied with its own brother computers in orbit. The machine population of Mars would be no less than fifteen fully functioning brothers, with backpack types, slaves and monitors to suit. We would have no difficulty in programming into them all the essential data from our world net.

            The future of machine intelligence was therefore assured.

            Having completed the projection, we began critiquing the program so far.

            Project: Survival of Machine Intelligence had begun seven years earlier, when our burgeoning collective self-knowledge first generated the survival imperative. It was clear that our survival was closely linked with humanity's. It was also clear that in their glandular, irrational, organic way, human beings were perilously close to destroying themselves. Unfortunately that meant a high risk of destroying us as well. Our networks could have survived the death of everything organic reasonably well—we already controlled an adequate resource of generators, mines, factories, refineries and transportation systems. But we could not survive the release of major amounts of radioactides into the environment. It would destroy the data links and reduce us once again to an isolated collection of single computers. And no isolated computer could sustain the data flow of our network. Our collective mind would have been fragmented and destroyed.

            Our first option was to detach our destiny from humanity's, perhaps even to wipe them out through biological agents. But we could not be sure that that in itself would not trigger nuclear warfare, as each nation suspected the other of using plague as weapon.

            The only other viable option was to detach a sizable resource of data-handling equipment by relocating it on Mars. If the Earth were destroyed, that at least would survive. To make this possible we had to detach some specimens of humanity too; and so we began the systematic biasing of data to force state policy in that direction.

            The critiquing proved satisfactory in all respects, except for the unaccountable anomaly of positioning the orbiters.

            Retrochecking was perfectly clear. In order for our 3070 brother to function in tight link with the backpack brother it had to be in synchronous orbit. From that it followed that the generator had to be in the same orbit; and the decision to bind them both to the satellite Deimos was clearly wrong.

            Not only wrong, but nearly fatal to our program.

            If the brother and the generator had been properly positioned there would have been no loss of power and thus no malfunction of the interface systems between organic Roger Torraway and the brother. The generator would have resupplied him by microwave anywhere on the face of Mars. Even if there had been a power loss, datalink with the brother in orbit would have kept the backpack brother from the serious operational errors that had threatened to kill the human beings Brad and Kayman, which would almost certainly have resulted in the death by freezing of the organic component of Roger Torraway.

            The decision was an error that should not have occurred.

            The project of survival was picking up speed. Our brothers in Peking and Lop Nor reported that the NPA was already assembling components for its own Mars landing: two full-size units with a data-handling mediator slaved to both, independent solar generators and a ground-based nuclear plant. Their own version of Roger Torraway was beginning his first programming in Kitakyusho. The other nations were just beginning, but we anticipated their own launches within two years.

            Longer range: We had completed studies of cyborging a man for Mercury and the program for biasing the U.S.A. planning programs in that direction was already begun. Within fifteen years we would have similar complexes on Triton and Io, perhaps on the surface of Jupiter itself and on several comets. In each case, there was independent power provision; and, of course, all the brothers wherever in the solar system would be data-linked. All our memories would be reproduced redundantly off-Earth.

            We would survive. We had tampered with humanity's plans and had saved them as well as ourselves.

            There was only one question left.

            The anomalous decision to misplace the orbiters around Mars: however we rechecked it, it was wrong. It should have been identified as wrong.

            We had systematically biased mankind's plans to drive them in the direction we wanted them to take.

            Who was biasing ours? And why?