The Trilisk AI (Parker Interstellar Travels #2)

by Michael McCloskey

  There was a word for this condition among Gorgalans—the annoying state of being in a planning phase, dropping into a sub-planning phase, then realizing the sub-planning phase required an execution section before the original planning phase could continue. Kirizzo was there now—he had to speed up his connection now in order to optimize the main planning phase. He thumped his last two legs forcefully in the Gorgalan equivalent of a curse. At least the execution of this sub-problem would not require physical action; that would have been an order of magnitude more frustrating and would have involved more cursing.

  Kirizzo contacted his ship using his own communications gear. It trailed the Iridar by many light minutes. It would never approach the home planet, for fear of getting the attention of the Bel Klaven. But it could connect to the Terran networks and get him the superior network access he wanted. Kirizzo had to pause and add a planning phase for accessing the Terran network without any of the usual accounting infrastructure a normal citizen had for identification.

  Finally, Kirizzo returned to his previous task of cataloging Terran methods of design and construction. He performed a review of materials commonly used, then expanded it to everything the Terrans could do regardless of expense. He moved on to design and control methods. Kirizzo entered all the information into his own storage and created a set of restrictions within which his design optimizer would have to work. The restrictions did reduce the problem space in which his optimizer could work, but it wasn’t too bad. The laws of physics, his available resources, and his end-product goals already formed a complicated maze to work within, so more restrictions would reduce the quality of the result, but the computation load didn’t increase that much.

  The planning phase came to an end, and Kirizzo shot into action.

  He constructed forty small devices, using methods and designs Terrans could have achieved. The devices lay before him in an organized grid as he proceeded component by component. He dared to improve things only a bit—perhaps five percent here and there—so his devices wouldn’t attract the attention of his enemies. Kirizzo didn’t know exactly how the Bel Klaven detected Gorgalans and their machines, but he felt his approach was sound. Most likely his enemies had fairly sophisticated methods, but given time and knowledge of the challenge, it would be easy to circumvent the danger.

  The Terran methods were so primitive compared to his own. Gorgalan technology was vastly superior, yet most of his race had been destroyed. Overwhelmed. They hadn’t had time to devise countermeasures, even against machines as inflexible as the ones the Bel Klaven had sent. Kirizzo was a Bel Klaven expert. He had become one in his fight to survive the machines that had hunted him. He knew that the secret to defeating the machines involved first understanding them, then outsmarting them. He believed the Bel Klaven made their war machines this way to prevent them from ever becoming a danger to their makers.

  The devices grew before him in ordered spurts. He fabricated the same pieces for all of them in batches as he went, using his own portable fabricator configured to emulate Terran materials and designs.

  At some point, Magnus came into the cargo bay and watched Kirizzo work. Kirizzo noted the healthy curiosity exhibited by his ally. As he finished a phase of the construction, adding power units to each of the forty devices, one of Shiny’s sensors routed an alert into his mind: it had detected the atmospheric vibrations of Terran speech. The message was diverted into the cognition layer which coated his long neural keel.

  “So, Shiny, I could use your help with the walker here. I want to adapt its power plant and these ingenious legs to my machine over here. I can’t interface with either of them, though. They use your generic computation blocks, but I can’t even tell if these blocks are working or broken. For all I know, I burned them out or damaged them just trying to scan the insides.”

  Kirizzo listened to the request carefully, yet he found the desired outcome to be largely unspecified. Kirizzo wondered how best to put power in the hands of Magnus to forge his own solution. That way, Kirizzo couldn’t be blamed for any failure in the outcome. He pondered how to reply; the Terrans seemed to prefer verbal communication even though their link devices were more useful. Kirizzo replied by vibrating one of its guardian spheres.

  “What is intended mission of the robot?”

  “I’m going to bring it with us on the next expedition. Telisa and I could use a safer way to scout ahead, and some backup.”

  Kirizzo considered the oddity of the situation. He attempted to adapt Terran technology in his drones, and Magnus was busy doing the opposite. He could understand the motivation to improve the machine but, given the current objective, he had to discourage the plan.

  “Potential disadvantage to using this technology: Homeworld destroyers will detect it, neutralize it. Optimal to avoid usage of other technology for the upcoming venture.”

  “Crap.”

  “Shiny does not excrete in the same manner—”

  “I’m expressing irritation. And before you misinterpret that, I mean I’ve been working hard to integrate this walker. Now you’re pointing out I shouldn’t do that. And I asked for help as part of the deal.”

  “Shiny suggest modular construction. Shiny-technology components could be swapped out for Terran ones as necessary.”

  “Very well. Help me out here. How should I interface with these legs? And this power source is amazing. How long will it last?”

  “Please wait.”

  Kirizzo had just examined typical Terran methods of design and physical construction. He now accessed several Terran robot manuals and examined their iconic methods of presentation, interface, configuration, and control. This dovetailed nicely with his own recent work on his probes. He simply had to expand his review into the area of manuals and protocols the Terrans had designed for themselves. Kirizzo created a model of the conventions used in the bulk of the manuals, fed it to a translator, and created a Terran style manual for the hardware of the walker.

  The Gorgalan computing blocks were harder. Their capabilities were so far beyond the controller used by Magnus, it would stretch the Terran’s imagination to understand. Kirizzo decided to simply provide a step-by-step guide for implementing the interface Magnus had created for the Terran machine on a generic Gorgalan computation cube.

  This took about fifty seconds, so at the conclusion of the work, Kirizzo simply transferred the Terran manual and the guide over to Magnus.

  The man stood up and froze. He didn’t say anything for a minute or so.

  “Thanks, Shiny,” he finally said. “This is exactly what I need. Wow, I have a lot of work to do, though.”

  “I will assist,” Kirizzo said. He assessed the materials nearby and entered another long planning stage. Magnus stood by, doubtless confused by the alien’s announcement in contrast to his sudden lack of activity. Some small segment of Kirizzo’s long, thin brain noted that Terrans differed as much from Gorgalans as the Bel Klaven did. He considered the notion that maybe their relationship would end as tragically as the Bel Klaven one had. After a minute of planning, Magnus stirred.

  “Thanks for that, but you’ve given me what I need,” Magnus said. “I’d rather fully understand what I create, so better if I take it from here. Besides, I have time to burn.”

  Kirizzo considered the words of the alien. The surface sentiment seemed reasonable; still, the possibility of a switch to competitive mode by either side required a model of hidden motives. Magnus probably feared the possibility that Kirizzo would install hidden control mechanisms if he had a direct hand in the construction of the device. But Terrans seemed to prefer extended periods of cooperation or competition over rapid switching; witness the continued cooperation between Telisa and Magnus. They were a mated pair, though; perhaps their race had optimized for slower switching within a family unit. If the Terrans did generally prefer extended periods of cooperation, why would Magnus fear the possibility that Kirizzo could be trying to get access to the robot? Probably because Kirizzo had revealed to
o much of his own race’s behavior patterns already. The Terrans were intelligent and they saw the possibility that Kirizzo would switch to competitive mode very soon, so they guarded against it.

  Could the Terrans have better models of Kirizzo’s behavior than he had of theirs?

  The one called Magnus worked a great deal on the scout robot. He seemed to follow the directions Kirizzo had provided, with a few modifications. The most notable deviation was in control software. This supported the hypothesis that Magnus kept his own software out of a sense of wanting the security of knowing more about how it would work—and knowing that Kirizzo wouldn’t have control of it.

  Kirizzo felt confident that should the need arise, he could take control of it anyway. It would be trivial to do when the robot was using the Gorgalan componentry; less trivial if using Terran controllers, but still very doable.

  Kirizzo returned a line of thought to their long-term behavior. The Terrans still seemed to be mated pair. Kirizzo hoped they would not stop the mission to produce offspring, as he felt certain that would slow things down. As it was, it was already painful to wait around while the Iridar made its way back to his homeworld.

  The Gorgalan tried to distract himself with some entertainment from the Terran net. He participated in various virtual games and explored many artificial alien environments posing as a Terran participant. These activities only accentuated the loss of his own Gorgalan network. He was able to spend time in artificial Gorgalan environments using his own hardware, but it lacked the depth provided by the others of his race. He was very alone now.

  He learned from their games and simulations that Terrans did have a concept of competition. However, perhaps like the Bel Klaven, their loyalties were more stable. The typical conditions causing a shift from cooperation were that one side or the other had feigned alliance in the first place, then initiated a shift when it was beneficial. These types were much more like Gorgalans. The shift was considered to be a bad thing to do before the other side did it, and there were dirty words for it: deceit and betrayal. It was clear that Terran mores demanded that loyalty remain in place long after it became detrimental to one party.

  Kirizzo formed the hypothesis that Terrans had ritualized their competition and permanently relegated it to the realm of sports activities and games. This gave them an outlet for their natural aggression, yet allowed them to remain in cooperative mode in reality for longer periods of time. If such were the case, though, it could hardly be a stable configuration. If a few individuals chose competition in the real world against the others, they would enjoy huge advantages that would quickly slingshot them past the others. When Kirizzo considered the elite who controlled most of Terran society on their homeworld, he thought that perhaps this had already happened.

  A key difference between the “evil” Terrans and typical Gorgalans was simply that the Terran betrayers often had never shifted to true alliance at all: they had pretended alliance all along. Gorgalans usually truly wanted alliance when it was asked for. They did not generally have any future betrayal in mind as a secret plan: it was just a given that the alliance would rapidly dissolve at some unknown time in the future when circumstances changed. A Gorgalan did not eagerly await or savor any switch to conflict, but when a situation changed, a Gorgalan changed fluidly with it and switched into competition more freely.

  At first Kirizzo thought this was favorable for him. Telisa and Magnus would very likely remain loyal to him until he decided the relationship should shift. That meant he could spend less time preparing for a harmful shift to competition and more time trying to get what he needed. But he could not be completely complacent: what if Telisa and Magnus were of the rarer, but deadly, “evil” variety? In that case, they might be feigning cooperation, even now secretly fostering an involved plan to switch to competition and seize a huge advantage. It was a risk-management situation, just as with his own kind, but the likelihood of a switch was lower on average.

  He concluded Gorgalan-alien interactions probably had different optimal switch points than Gorgalan-Gorgalan relations. It had all gone so wrong with the Bel Klaven...

  Kirizzo remembered the beginning of the war. Several powerful Gorgalans, initially in opposition to each other, had decided to cooperate to forcefully compete with the Bel Klaven. Working together, the suddenly aggressive Gorgalan alliance seized several resource-rich Bel Klaven planets.

  The Bel Klaven had never experienced the rapid switch of behavior inherent in Gorgalans. Almost overnight, their greatest allies had turned into powerful aggressors. The Gorgalans had always remained vigilant, hedging against such a turn in the Bel Klaven, which had never come.

  Instead, the Bel Klaven shored their defenses and bided their time. When the Gorgalans came back to offer cooperation again, they were rebuked. When other Gorgalans unrelated to the attackers offered friendship, they too were rejected. The Bel Klaven blamed all Gorgalans for the actions of the powerful alliance. In fact, it had been a round of backstabbing back home by powerful Gorgalans such as Kirizzo that had ended the offensive against the Bel Klaven and brought the alliance back home to protect its interests on the homeworld.

  Kirizzo hadn’t participated in the attack on the Bel Klaven, but he had profited greatly from it. The distraction had allowed him to gain a lot of ground back home.

  Those gains were obliterated when the Bel Klaven revenge fleet arrived and dismantled the planet by force of arms. If Kirizzo hadn’t been far away, trapped in a Trilisk ruin, he would have died with millions of other Gorgalans.

  Telisa shared his interest in the Trilisks and their technology. That was both good and bad. Though she might assist in procuring new technology, she would also be a competitor when it came time to split the spoils. Fortunately she’d been easily satisfied by small items here and there. She didn’t have the grand designs Kirizzo did.

  That was almost certainly because she was ignorant of exactly what the Trilisks were and what they did. Kirizzo had made it much further on that front, despite the fact that the Terran homeworld was itself a Trilisk outpost. After examining their history, he was sure of that. If Kirizzo had more of a sense of compassion, he would have felt sad for them. As it was, he considered it a study in failure, but whether it was a failure of the Trilisks’ plan to help the Terrans, or a failure of the Trilisks to seize Terra for themselves, Kirizzo hadn’t quite figured out yet.

  Telisa and Magnus would have to go into his house and retrieve the industrial seed. The odds of their success were hard to calculate, but it was easy to see that they would have a better chance than he ever could. The seed was key to Kirizzo’s chances to start over again. Without an industrial base, he would be doomed to spending the rest of his life wandering about in Trilisk ruins, hoping to get bootstrapped again to where he was before. It would be immensely frustrating.

  The Gorgalan alliance’s decision to attack the Bel Klaven had turned out to be a terrible one. Kirizzo decided to remain in cooperation with these aliens, at least until he could regain what he had lost.

  Chapter 7

  “I assume you’ve started the search for them. Any ideas yet?” asked Arlin. Relachik and Cilreth sat in the galley of the Vandivier as Arlin hovered in the doorway.

  “It’s too early to say. I don’t think they came back to Earth, though.”

  “I agree, though I’m only speaking from the gut,” Relachik said.

  “It will take me a while to get set up. One thing I can tell you is, we’ll know in a day or two if they’re serious about hiding or not. If they’re being sloppy, this will be quick and easy. Otherwise, we may have to wait for them to screw up.”

  “I heard the UNSF uses artificial intelligences to search for things, and people,” Arlin said from the doorway. He dodged in and grabbed a snack in the tight space before moving back out to the door.

  Cilreth nodded. “Maybe so, but that’s not as useful as you might think,” she said.

  “Are you flattering yourself or insulting the government
?” Relachik asked.

  “Neither. Just talking about diminishing returns in intelligence,” Cilreth said.

  “I’ve heard that many times, but I don’t follow it exactly,” Arlin said.

  “We have artificial minds that do basic computations millions of times faster than we can. But that’s still not enough, because as the number of facts rise, meaningful interactions between them explode quickly, forming a mountain of possibilities so steep that even something much faster than us still can’t work through it all. The machines can make it farther up the curve than we can, to be sure, but they’re only about a third smarter than the smartest of us.”

  “If it’s so much faster, that means each second is a long time for it to think,” Arlin said. “If you put me in a room for a million years, I could solve a lot of problems.”

  “Given a million years you could go through a lot with a small set of facts. But given a large set of facts, the permutations of all of them, their causes and effects, their associations... the number of possibilities explodes rapidly as the fact set grows. It’s a combinatoric explosion. Considering the interactions of ten facts, possibilities, or events is much more than twice as hard as considering the interactions of five things. A mind with machine memory, incredibly fine senses, the ability to think about a hundred things at once, incredibly fast net connections, and everything else a large AI has, is confronted with millions of facts every microsecond it’s alive. It has to wonder whether the third microbe from the left on the rightmost ceiling tile on the last row has anything to do with the murder of Mr. Mustard.”

  “Colonel Mustard. But we discard useless facts like that,” Arlin pressed.

  “It may be useless or it may be the only remaining microbe of the disease that killed him. But yes, you’re right, part of intelligence is about figuring out which facts to examine and which ones to discard. The only way to control that explosion is by aggressively culling facts that aren’t important. Humans do that all the time since we can only handle a few ideas at once. But care is needed: discard one fact necessary to solve the problem and you’re stuck. And if an AI culls its fact inventory all the way down until it’s aware of only the things a human is aware of, then it’s only as smart as a sharp-minded human, though somewhat faster. It threw away all the extra facts that could have made it godlike. Somewhere in all those facts are chains that could be used to make amazing deductions, but the power it takes to analyze them rises exponentially with the number of facts.”