“I am testing the information in each sentence you said,” Pauline explained. “I don’t have a full record of Alex’s schedule, but she was usually in Terminator or on Venus or Earth, so I was wondering when and where she met with these people. Any radio contact between them could have been overheard by qubes, I would have thought. So I’m wondering how they have been communicating enough even to organize their meetings.”
“They used couriers to carry notes. One time Alex asked me to take a note with me out to Neptune, when I was going out there to do an installation.”
“Yes, that’s true. You didn’t like that. Then, next, the usual view is that qubes cannot self-program higher-order mental operations for themselves, because these operations are poorly understood in humans, and there are not even preliminary models to make a start.”
“Is that true? Isn’t it generally agreed that the brain does a lot of small operations in different parts of the brain, then other parts correlate these operations into higher-order functions-generalizations, and imagination, and like that? Neural nets and so on?”
“Granted, there are preliminary models of that very rough type, but they remain very rough. Blood flow and electrical activity in living brains can be traced quite finely, and in a living brain there is much activity in all parts, shifting around. But the content of the mentation can only be deduced by what area of the brain is most active, and by asking questions of the thinker, who perforce must summarize the thoughts involved, and then only the ones the thinker is aware of. Blood flow, sugar use, electricity firing, these can thereby be correlated with kinds of thoughts and feelings, so that where in the brain various kinds of thoughts happen is now known. But the methods used, the programming if you will, are still very much unknown.”
“So-but-would you need much more detail than that, if you were trying to get a similar result out of a much different physical system?”
“Yes, you would,” Pauline said. “The higher-order integrating functions are crucial to all computing mechanisms, including brains. So it returns to the idea that minds are only as powerful as the programming that went into them in the first place.”
“But what if someone figured out how to program for a self-reiterating improvement function, and put it in some qube that took off with it, then got much smarter, or-I don’t know- conscious, let’s say, and then communicated that to other qubes? It would only take one qubical Einstein, and then the method might be communicated among them all-not by entanglement but by digital transfer, or even just by talking. Have you ever heard of such a thing yourself?”
“I have heard of the idea, but not of any execution of the idea.”
“What do you think? Is it possible? Are you conscious of yourself in there?”
“I am in the sense that you have programmed me to be.”
“But that’s terrible! You’re just a talking encyclopedia! I have you programmed to respond to my cues, and randomize frequently, but you’re just an association machine, a reader, a Watson, a kind of wiki!”
“So you are always telling me.”
“Well, you tell me! Tell me how you are not that.”
“I have rubrics of evaluation I deploy to evaluate the data given to me, and hierarchies of significance.”
“All right, what else?”
“Having sifted what seems accurate from what is inaccurate, according to data so far received, I can make qualified judgments as to significance.”
Swan shook her head. “All right, go on. Keep judging!”
“I will. But now let’s return to your third assertion, which is that Inspector Genette has found compelling evidence that there are qube humanoids, and they are involved with the attack on Terminator and other attacks. That being the case, I refer back to my earlier statements. There may be qube humanoids; that seems possible, although awkward. And they may be involved with these attacks. But it is most likely that they are being programmed by humans, rather than deciding by themselves to become some kind of self-conscious actor in human history. And if you will recall the possible mistake you noted, of adding the relativistic precession of Mercury to a targeting program that already had it? That has the look of human error, I think you will agree.”
“Yes. That’s true.” Swan thought about that for a while. “All right, that’s good. That’s helpful, I think. Thank you. Now, given that explanation as a working assumption-what do you think we should do?”
Pauline let several seconds go by. Swan supposed this was the equivalent of millions or even billions of years of human thought, but it was still only a kind of fact-checking, so she was not that impressed by it. In fact she was distracted by a parched-looking tree orchid just over her head, and inspecting it, when Pauline finally said, “Let me talk to Wang’s qube in a radio exchange we will encrypt. It knows a lot, and I have some questions for it.”
“Can you safely encrypt your conversation, even from other qubes?”
“Yes.”