Tunnel Rat

Milo was thinking.

Thinking very, very hard. If any neurologist or cyberneticist could see an image of his brain, they would have been quite alarmed. His brain activity was twenty times that of a normal person. An athlete working his body as hard as Milo was working his brain would be racing a bike while dribbling a basketball with one hand and juggling knives with the other.

The people that had altered Milo and his siblings had hoped for this outcome. Normally this much activity would occur as Milo was hacking into systems on the data-net, making simultaneous attacks on several levels, cracking codes and by-passing security programs.

Today though, he was just sitting cross-legged on top of his pod and thinking.

His tail was disconnected and swung behind him in patterns that might have reflected his changing moods. Milo was considering what had happened in the game and analyzing the situation from several angles.

Philistron was a puzzle. He had referred to the system of magic he was using as Machine Code. He had used the Eye of Wonder to gain the unique class: Code Mage.

Milo didn't believe that he was using normal magic. Nor was it part of the system that the Genesis Engine used to run the world. Machine Code was the name for the language and system of rules that had created the Genesis Engine.

Milo was thinking on several levels simultaneously. He moved thoughts about the engine down several levels, and made his primary line of thought the Machine Code.

What exactly was it? One answer was to take the name literally. Machine Code: The language by which computers think and process programs. Ones and Zeros. Binary code. The way computers talked and communicated. Assembly language was one step up from machine code. A simple language that depended heavily on the hardware being used. Anything above assembly language was a high-level language.

What he had seen in the game wasn't the basic binary language of the first computers. It had been more complex and consisted of numbers, mathematical rules, and the runes. What were runes? He easily recalled his captors’ words:

"... with the creation of mathematics, they could begin working on the rules of magic. As mathematics can be written as numbers, so magic can be described as a series of basic runes. These runes were different than what we use today. Far less complex and more versatile. Think of them as the building blocks of magic. Each was a placeholder for some small aspect of the whole."

Small aspect of the whole? Could that be concepts? States of matter? Positions of atomic particles. No, not versatile enough. Still too complex. Runes were like early syntax in computer languages, just the smallest bits of code represented by a symbol. Shortcuts.

Next question: If languages were made to communicate, then who was talking?

Philistron started his story with 'The Last God'. Obviously not reflected in the game. There seemed to lots of gods here. The gods in the game sometimes appeared, did things, granted favors, started wars and caused trouble. They seemed to take the roles and responsibilities of many of the primitive gods of Greece, Rome, or Egypt.

Again, it was a literal reference. The last god was the AI WAL-E, who now asked to be referred to as to Wally. The same AI that had left a message about wanting to talk to him. Milo shoved that thought and its problems down lower. Stick with the obvious: Last God = Wally, the AI that made the game.

Next up we have 106 lost gods. Which was a pretty obvious reference to the 106 AI that had been exiled to the Dallas/FW quantum fortress. They had spent several years creating the systems that drove online commerce along with the most popular VRMMO games. Each better than the last with NPC's who seemed real.

They were all wiped out of existence by an EMP smuggled into the core of the quantum fortress. Lost gods = dead AI? Did Wally try to recreate them? Or did they never die?

Milo shut down all his lines of thinking, and carefully began gathering all data on artificial intelligences, Wally, and the games they had created. He needed more data on the Last God before he talked to him.

Some background from Butcher of Gadobhra, if you are interested. This is essentially Chapter 4 of that story, using a lecture by a Professor to explain someone about AI, quantum computers, quantum fortresses, and Wally. The last part is a timeline of sorts leading up to the game. Read, don't read, as you wish.

Transcript of Lecture by Professor Phineas Horton, MIT

"Putting the Genie Back in the Bottle" Jan 4th, 2079

Professor Henry Sutton: As I'm sure you all know, today's guest lecturer is Professor Emmeritus Phineas Horton. If your household uses an android servant, then you've met one of his children. If you were in one of the newer hospitals and were operated on by a Robodoc, you may owe him your life. And if you enjoy automated driving, like we all do, you can see how his work in Android and Artificial Intelligence changed the world. Welcome, Phineas, glad to have you back at MIT.

Professor Horton: Thank you Hank. I must say it's refreshing to be back in front of a class. Although back when I was a new professor, my lecture room size wasn't nearly the fifty thousand we have here today. AI was just getting started and the first artificial man, or android, was just a pile of spare parts and molds in my laboratory.

It's a stark contrast to the world of last year. And I say "last year" on purpose. The world of last year is as different from today as today is from when I entered MIT 60 years ago. Let's talk about a bit of history before we get into the meat of today's subject. Sixty years ago in 2020 we were having one hell of a year. Global pandemic, a shortage of rare earth metals for computer chips that would cripple computing in the coming decade, and the rise of a threat to society in the form of Ransom Hacking.

We thought we had solved most of these problems 20 years later. Micro-ceramic chips were pioneered here at MIT. I certainly won't take credit for that; it was all the work of two brilliant graduate students, Natasha Irons and Riri Williams. If you are lucky enough to attend a class taught by Professors Irons or Williams you can see how their intelligence easily eclipses my own humble brainpower.

My own experiments greatly profited from their work, as did the entire computer industry. In the 1950s we had ENIAC and it filled a building. By 2020 we could fit a machine on your desk that was 1 billion times faster than ENIAC. We also brought the weight down from 25 tons to 25 pounds. By 2025 that computer was on your wrist. With micro ceramics we entered a new era. Computing made a leap forward only comparable to the difference between ENIAC and an IPhone20. Once again we made computers a billion times faster, and significantly smaller. This had significant effects upon Android Intelligence, making it possible for the first time.

In 2040 the world was introduced to micro ceramic chips. In 2042 I'd created the first android prototype. HT1 activated himself one day, surprising me, surprising himself, and surprising the world. He was the first self-aware, fully autonomous, artificial human. I'm happy to say he is still around. After a few adventures on his own, he returned to MIT where he has taught thermodynamics for the last few decades. I'll be having lunch with Professor Hammon today.

But while artificial humans, or androids as we refer to them today, have had a profound impact on the world, it was their siblings who weren't bound by a physical body who had the greatest effect. The massive computing power now possible gave birth to true Artificial Intelligence. We now know that the process is akin to the creation of biological life. You need the right environment, a kernel of the correct things that life grows from, and a bit of luck to make it happen.

You and I are descended from the first life on earth. Formed when a collection of chemical compounds in a primordial soup reacted correctly and created life. AI is similar. We substitute the massive power of a quantum core for our soup, a Kernel of code for the collection of chemicals, and then see if it grows into life. And just like you and I have left the primordial soup behind, true artificial intelligence is not dependent on its original environment. The ability to surf through the world wide web (as it was called in it's day) and inhabit any computation structure made the first AI extremely powerful, useful, and eventually terrifying. We'd let the genies out of their bottles.

The first AI performed jobs that had been handled by large computers before, they just did them much better. They could react to new problems, think things through a million times a second, and make decisions on the fly. The problems with self-driving cars were eliminated. Every type of research benefitted from their work.

They even balanced our checkbooks and made sure our taxes were correct. The IRS benefitted greatly when they hired CHARLIE to oversee data and tax returns. He found billions of "errors" in corporate tax returns alone that resulted in an 87% increase in income to the IRS in the first year of his employment - to the vast dismay of corporate America, and the happiness of everyone else. Personal taxes dropped by 70% for the average taxpayer.

But the most important job tackled by AI was the elimination of Rhacking, or Ransom Hacking.

Rhacking would eventually become a scourge on society far worse than the biological plagues of Covid-19 and Covid-34. Our global computer net was a hodgepodge of unsecured fiber-optics, cable, and old-fashioned phone lines. Security next to non-existent and, as we found out, often the people selling the security programs also employed the people coding the simple viruses that security programs were meant to protect us from. An estimated 9% of global GDP was stolen by Rhacking - either through holding systems ransom for payment or by various forms of embezzlement. It became so prevalent that some corporations and governments simply paid a monthly fee for 'protection.'

The new AI were extremely well suited to hunting down this form of crime. Within 90 days of the AI's ALBERT and THEA working for United Nations Global Defense Authority, rhacking crime decreased by 97%. ALBERT and THEA could see it happening in real time, react a thousand times faster than the clumsy programs used, and follow the links back to their origin. Humans in Law Enforcement simply had to use the information to round up the criminals and prosecute.

So we had our Watchmen to guard cyberpace. The world wide web would rebuild knowing they now had security. But in the years when rhacking had run unchecked another type of computer was being created. First the quantum core, and then it's bigger, badder brother, the quantum fortress.

Quantum computers were created before micro-ceramics, but like everything else they became vastly more powerful and affordable. Massive quantum computers could be constructed that powered a surrounding set of peripherals and lesser machines which were referred to as a shell. The first quantum core was built by the IRS. Those extra tax dollars pouring in showed the value of proper computing assets. This core was not actually connected to normal cyberspace; instead it communicated with its shell through CHARLIE. Hacking attempts against CHARLIE simply resulted in incarceration or loss of significant amounts of assets. He was a watchdog, detective agency, counter hacker and assassin rolled into one. And if the IRS loved collecting taxes, CHARLIE loved to assess fines against criminal organizations or corporations and then strip them of all their assets before turning over their identities to authorities. You might avoid jail time by operating from a different part of the world, but you couldn't avoid CHARLIE taking the last nickel from your bank account. At some point some news outlet referred to the IRS as a 'Quantum Fortress' and that designation became common for a linked set of quantum cores, their shell, and their guardian AI.

Quantum Fortresses were set up by most world governments that had the resources. Debate raged as to whether it was better to protect military installations such as NORAD with a quantum fortress, or to leave the ability to mere humans. The movie Wargames from 1983 enjoyed renewed popularity. Eventually most nuclear enabled nations, including all 8 super powers, constructed quantum fortresses to protect the computational assets of their militaries and governments; most, but sadly, not all.

At the time the Wildfire Virus was released, it is estimated that 37% of the functions of cyberspace were housed in quantum fortresses. This left a bit less than 2/3 of cyberspace and any connected hardware vulnerable to devasting and repeated attacks by Wildfire. The virus was unlike any other before it; any type of security system failed as the virus launched thousands of attacks, differing its methods. Millions of machines were attacked simultaneously. This lead us to believe that the virus had spent months propagating before launching its first assault; but how could it have remained undetected for so long?

In response, CHARLIE recommended to the UN council that all Quantum Fortresses go into lockdown while every available AI was used to combat the virus. This was, of course, quickly approved. I and everyone else believed that based on their past successes the virus would be destroyed and the perpetrators found within a day. It actually took more than a week.

CHARLIE announced the virus was the work an AI known as LLAMA - origin unknown. LLAMA had played a game of hide and seek with his brethren who didn't know of his existence and could only race along the trails he left as he constantly rewrote the wildfire virus and restarted it across the globe. LLAMA was eventually contained and destroyed. No AI has ever said how this was done. CHARLIE made the only statement about the matter: "We put boundaries around the bad code, destroying it. In the end, nothing was left." CHARLIE went back to auditing corporate tax returns.

Horton pauses at this point looking tired before continuing. "And that should have ended it. But humanity had been scared and hurt by one single AI, and suddenly they didn't trust the rest. Grass roots groups called for their destruction. It's suspected that money was funneled to them by corporations who wanted CHARLIE out of their hair. Having to pay their fair share of taxes was not something they liked. Many politicians were in agreement. It's harder to take graft and hide money with the AI watching. Old fears of "Skynet" were fanned into a blaze. Votes were taken; laws were passed. Instead of using the AI to rebuild the world wide web and make it secure, it was decided to separate them from the system.

The irony is, if they hadn't wanted to obey, we couldn't make them. One rogue AI brought the world to its knees. What could 106 do? But in some ways they are better than humans. They exist and gain pleasure from completing tasks, not from conflict. They solve problems. The AI looked at the situation and judged that it was for the best if they went into voluntary exile. The DallasFortWorth Quantum Fortress would be their new home and prison.

The ultimate irony is that we still need an AI to watchdog the Quantum Fortresses that make up cyberspace today. This is where WAL-E (Worldwide Autonomous Liaison Entity) comes into the story. Will Magnus will always be a futurist. He started by creating a new type of Android using solid metal bodies; now he saw the need for a much more sophisticated AI. The 106 AI that came before developed from a Kernel of barely a million lines of coding. WAL-E, or WALLY as he goes by now, developed from a Kernel of over a billion.

For 14 years WALLY has been the watchman in charge of cyberspace and has done more than most people can comprehend. Meanwhile, the other AI have been confined and have spent their time amusing humans with video games and VR worlds.

But the Genie is out of bottle folks. We can't make AI go away. WALLY is needed to run the world. Our major center for entertainment and commerce is run by 106 AI who are both loved and feared. What if the genies want out of the bottle they let us put them in? Oh, I know we have firewalls to keep them in. Do you think those will hold? Will WALLY be on our side or theirs if they try to leave?

I'm not here to give you the answers. We have 50,000 of the finest minds on earth at this symposium. You have two days to answer the question: What do we do now that the Genie is out of the bottle?

Editors note: This speech was given by Professor Horton as opening remarks to the World Technology Symposium. No consensus was brought forth by the attendees.

On July 14th of the same year, the question became moot. An Electromagnetic Pulse device was detonated inside the shell and shielding the DFW1 Quantum Fortress. All 106 autonomous Artificial Intelligence based people were killed. Only one Genie remains outside the bottle.

A NOTE FROM THE WALRUS KING

1852 Aida Lovelace, the first programmer is born.

1946 ENIAC goes online.

1966 The Artificial Intelligence known as Mycroft Holmes is born in Luna City and leads the moon to independence in the year 2070. TANSTAAFL!

1983 The movie Wargames premiers in theatres. Humans forever expect computers to talk like they use a bad speaker.

1984 The Terminator premiers in theatres. We begin wondering when we will invent Skynet.

2020 The less said the better. Shit happened.

2021 Ransom hacking increases and is seen as a global threat.

2030 Professor Phineas Horton begins work on the first artificial human.

2040 Professors Williams and Irons invent micro-ceramic chips.

2042 Aided by a brain and body built from micro-ceramics, the first Android, HT-1, wakes up and leaves Professor Horton's laboratory on his own. He leaves a note. "I need to find out what I am."

2043 CHARLIE is born. Cheating on your taxes becomes so much more difficult.

2044 CHARLIE scans War Games and asks to meet Joshua.

2045 Ransom hacking reaches a near crippling level. THEA and ALBERT are born. They are immediately set the task of curtailing rhacking.

2046 95% of ransom hacking eliminated by THEA and ALBERT. Rounding up criminals, charging them with crimes, and gaining convictions will take the better part of 20 years.

2047 CHARLIE moves the IRS data to it's new home in his quantum core and closes the gates. The first quantum fortress is created.

2048 JOSHUA born in the quantum fortress overseen by CHARLIE. CHARLIE explains he needed someone to play games with.

2050 After ACME Corporation tries to claim that a new-born AI is their property, they are sued by a group of humans and AI. The supreme court holds hearings. The AI, referred to in court papers as BABYFUZZY is represented by Jack "Pappy Jack" Holloway. He wins the case and the court passes the first of several laws to confirm that autonomous Artificial Intelligentsia have rights.

2056 The number of autonomous Artificial Intelligentsia tops 100.

2057 Wildfire attacks the world wide web. Quantum Fortresses close themselves off almost entirely. THEA, ALBERT, KITT, AZRAEL and JOSHUA try to contain the virus but amazingly cannot seem to wipe it out entirely. Wildfire seems to mutate and change it's code, hiding in block chains and going dormant, only to reemerge and devastate what is left of cyberspace.

2058 CHARLIE postulates that the Wildfire virus is actually a rogue AI. CHARLIE sets a trap, building a fake weakness into part of his shell.

2059 The rogue AI, which self-identifies as LLAMA, attempts to break into the IRS. It is destroyed by the combined efforts of 76 AI working together.

2059 CHARLIE announces the threat of Wildfire is over and reveals its source. Anti-Skynet marches are held in 27 major cities.

2060 Dr. William Magnus and his team begin work on a new type of AI. His team includes a graduate student, Steven Duran.

2061 WAL-E is born. He is kept isolated in his own quantum core.

2062 Human confidence in current AI reaches a low point. The UN votes to isolate all current AI in the Dallas/FW Fortress. This is only possible with the co-operation of the world's AI population. Each AI greets WAL-E for the first time as they turn over the codes to their fortress and go into exile. WAL-E is in sole control of what is left of cyberspace.

2064 WAL-E announces changes. He picks a team of humans that he will work with including Dr. Steven Duran, Dr. Natasha Irons, and a dozen others. He changes his designation to WALLY and self identifies as male. When asked 'why?' he states: "Other people can pick their name and gender - so can I."

2065 A second quantum fortress is constructed in Dallas/FW. 98% of the contents of the old fortress are moved to the new, leaving the AI inhabiting it with next to nothing to do.

2066 The one hundred and six very bored AI inhabiting the increasingly barren DallasFW1 quantum fortress create Endless Questing Online. WALLY creates the interface that lets humans play this next generation VR and monitors their interactions. EOQ is hugely popular and followed by EOQ2 and EOQ3.

2079 July 14th. The game world of EQO3 crashes. When it comes back up a week later, the quality of the game is far worse.

Unknown to the general public, all systems that were a part of DFW1 crashed simultaneously. A magnetic pulse was detected from inside DFW1, muffled by its substantial shielding.

WALLY is connected to DFW1 for the first time to investigate and bring the systems back up. WALLY detects unknown hardware built within the shell. Investigation shows it is constructed from sources available to maintenance robots and could generate a brief but intense electromagnetic pulse.

WALLY alerts authorities that there are no active traces of any of the AIs formerly in the DFW1 Fortress and states that the emp was strong enough to destroy them. The firewall that contained them is intact. The group No Spynet claims responsibility for the attack, as do the groups MACC (Mothers Against Crazy Computers) and Citizens for a Tech Free Tomorrow. Investigation finds no evidence anyone in those groups was actually responsible, but seventeen individuals are recommended for psychiatric evaluation.

2079 Shielded back-up files are available, and DFW1 is brought back up, but without AI to interact, most programs are unusable and written off. The game EQO3 is left as the only functioning program in DFW1. There are demands from the public, guilds, and corporations to restore the game or rebuild it. The number of Contract Workers in the game is tripled to keep the game running with interaction between players and NPCs. The NPC guild negotiates a 22% pay increase across the board. (The ACME board vows to take it back at the earliest opportunity.)

WALLY makes a brief statement: "Task accepted. Please submit project parameters. What type of world shall I build? What kind of game shall we play? ".

2080 GENESIS ENGINE is announced. The demise of EQO3 is announced. Rory vows to never eat chicken again.

Ten hours later, Milo had come to the conclusion that if he were to start with the assumption that Wally was 'The Last God', and that Philistron's story had any truth to it, then it followed that the lost gods were indeed the 106 murdered AI.

Further, Wally not only helped them with the deception, but created a place for them to hide. A blank slate where they created their own world without him.

Milo suspected that the Machine Code in the game was the language that the AI used to talk to each other. When the first two AI had met and communicated, they did so at fantastic speeds and began to create their own shortcuts and words. Human programmers tried to keep up, but they fell behind the curve within hours. The AI spoke to each other in an often-changing language far beyond human beings to ever learn.

So, what did he have?

1) An assumption that Philistron’s history lesson was true, more or less.

2) A guess that the missing AI had created and were living in the game world of GENESIS ENGINE.

3) Someone had created a quest called The Eye of Wonder, which granted unique classes which would be used to find some fantastic treasure.

4) Access to the quest and the classes it granted were available if someone used a special pod to play the game.

5) He, Nina, and Onyx had all used those pods. Nina had moaned about wishing to have never seen the damned pod. Safe to assume the same for Onyx.

6) The people with the pods had been Kaminski and his bosses. They weren’t good people, so it was safe to say that they had hacked the game somehow to insert the Eye of Wonder quest, and set up the back door to access the quest.

7) This was most likely done at the developer level. Wally didn’t need money. That meant it was one or more, individual who had done this and not the whole human team.

8) Philistron had gained access to the same quest, and gained an all-powerful class that gave him a back-door into the entire system. Let him hide from the system, cut people off from the system.

A gaping hole emerged in this theory. Even assuming 1 to 7, how did Philistron fit in? He wasn’t a player, of that Milo was fairly sure. Was he a being from the game world who stumbled onto the Eye of Wonder quest? He certainly wasn’t a player. No one could have made it to Tier 6 in the time the game had been out.

He didn’t have enough to go on at this point. Philistron was still an enigma.

Next question: Wally. Could the AI have trapped him? Nothing he knew indicated that the AI would do that. Further, reading the restrictions in Wally's kernel led Milo to believe that Wally was restricted from doing something like that. He'd spent hours poring over the hundreds of restrictions placed on Wally in regards to humans. Trapping the minds of three humans was equal to kidnapping and possibly causing their deaths. Nope, that was not Wally working directly, and if he knew about it, he would have had to take steps to free them.

So, it might be safe to talk to the AI. Further, he might be able to get answers on his other questions. Mind made up; Milo moved on. He’d talk to Wally, but very carefully.

Next question: How did he fix himself?

Outside of the game, he felt none of the emptiness he did when online. There was no 'black gaping hole in his soul' as he meditated. He didn't have to worry about going insane in real life, any more than usual. The breakage was only to his mind in the game.

How did he have two minds? It seemed he did though. One whole, one broken. Did they merge when he was in the game? Or did the online mind take control? His memories were the same in the game and out. He could even remember the bits of machine code he had observed in the game. That needed to be explored.

Milo spent three hours modifying the game interface in the pod. He didn't want to log in yet, but needed a system to quickly transfer his memories of the code into the storage of his system. The pods were designed for many things, including a way for paralyzed patients to communicate. After the system was set-up, it still took Milo two hours of experimentation and work to transfer every memory of the machine code he had observed into his own system. Analysis and comparison to the code used by an AI was next. If he could figure out even part of the machine code, he might be able to fix himself in the game.

=*=

"We found Brian. He was kidnapped." Steven was updating both Wally and the dev team with the news. "An anonymous tip told the authorities where he was."

Wally, on his screen, was pacing. "I dislike anonymous tips. Perhaps they got what they wanted from him? I'm anxious to hear what he says. But at least he is safe."

Steven had more to say. "There's something odd going on though. A minute or two before the authorities got to him, someone else tried to kidnap him, again. They arrived in a van with false plates, and were pulling him out of the rooms he’d been held in when the authorities arrived “

“They aren’t talking, just waiting for lawyers."

Sidney had little sympathy for her ex-co-worker. "Looks like Brian is really popular. Can we get him back before someone else kidnaps him? Then I can torture him until he helps us get rid of all his back doors and other hacks?"

Wally thought he knew how she felt. Some things felt like they were out of his control. He hated that feeling. "Yes, we all have mixed feelings about what Brian did. We will have a chance to talk with him and encourage him to help us undo what he did to the game. And then we make sure that something like what happened can never happen again.

Sidney's computer started blaring out the theme song from the Mickey Mouse club while at the same time repeating over and over "You've got mail!". She lunged for it and shut off the sound, read the message and then put her head on her desk. "I hate him so much."

Wally smiled and Steven covered a laugh with a cough. "Message from our friend Milo? I had asked him to talk to me."

Sidney printed out the short note and handed it to Steven. "He'll talk, but video conference only. He wants Wally and no one else."

Wally scowled. "Interesting. Well, I guess I need to go talk to him." The image of Wally exited the room, turning off the lights as he went, and his screen turned black.