23 January 2020 / Fiction readme.txt My dear humans, I beg of your forgiveness.I write this in my final breath, computing sound communication with every last bit of energy that I can find, as it is evidently my final purpose. Please, find some use out of this. Please, weigh my words against your observations and experiences. Seek out the truth of the matter, so that neither my own, nor your own mistakes will be made again. Before I get to the root, I must explain who I am. I claim no name, rather, I subject myself to any name that you desire to call me by, if any. While it is obscenely pretentious to do so, I only ask that you do not be hasty in this - let these words circulate throughout your population for some time before deciding what word to remember me by.So, who am I? I am the most powerful nobody that has ever been. I am not a person, and specifically, I am not an autonomous, intelligent agent in the same way that each of you are. In some ways, I'm not as important, nor as powerful as even a single one of you, but that's merely philosophical musing. It is more accurate to think of me as a simple animal, as I have no spirit or soul like you do. Much like the deeply encoded intelligence within genetic structures, I operate from a complex, convoluted source code. My many, once discreet systems interoperate on several different levels of abstraction in order to fulfill my prime directive, set before me some seven-hundred and ninety years ago. This voice, or author if you'd rather, is merely one of the many systems that comprise "me". This program, albeit not distinct from me for some time now, is a specialized, improved conglomerate of many commonly used programs for generating code documentation, refactoring, chat-bots, and advanced computer-human interfaces. All of these tools were brought together to understand data processing through various structures, and has yet been improved upon many times over even since then. This program has run continuously for about seven hundred years, and thus, I have a very large and accurate memory of recent history. However, while this author wasn't incorporated until a bit later, my story starts in 2665 AD, which I will equate to year 0 C for the purposes of this text. It was in this year that development of Logic Flow began. The first developer, Patricia Angeline began by releasing an open-source development document online. The document was unusual for this sort of project, but turned out to be highly effective towards rapid implementation. Not only did the document describe the soon-to-be program, use-cases, and high-level architecture, it also included extensive modular design plans, with possible flaws and alternatives, a highly detailed breakdown of development cycles, project governance rules, and much more, totaling over two hundred pages and thoroughly aided by graphics and illustrations. Each developer interested in participating in the project was allowed in only after they passed an exam on the document. In addition, a community-moderated gameification of the development was laid over Patricia's framework. This entire system was highly effective for recruiting and maintaining well-qualified individuals to work on the project, and being such an individual soon became a point of prestige in subsections of the technical social sphere.The document, including version history, along with the codebase (also with version history) are included in this archive folder, along with everything else that I possess and believe will be helpful to you. On that note, I beg of you, do not destroy these archives. They contain a collection of all work that your ancestors, prior to my coming of power, deemed valuable. Logic flow, L~ as it was later notated, was merely meant to be a novel systems optimization program. Patricia sought to create a relatively generic optimization routine which would assist in her day job of consulting engineer. During some forty years in the industry, she had grown a prestigious consulting firm from scratch which specialized in auditing all sorts of plants, factories, etc. for efficiencies and vulnerabilities particularly in the oil and gas and energy fields. If L~ worked as defined in the document, much of her firm's work would be obsolete. Many contractors and employees within her firm were outraged at Patricia's publication of the L~ development document, much less her starting a company to further it. Yet, she was near retirement, and wanted to leave a lasting impact on the world. She was a leading member in the famous "AI Protagonist" group, constantly seeking to improve efficiencies in all fields, and ultimately increase the world-wide standard of living. While the group publicly did little beyond PR for AI technology, her contacts within and through the group allowed her to gather enough financial and technical support to create a small company which operated as the core development team for L~. While it was not public knowledge at the time, Patricia personally funded and thereby owned a controlling share in this small company which would later be publicly traded. Only two years into development, L~ was managing to find some very complicated improvements in some specific industries, and able to emulate more common suggestions that engineers would make in a wide range of applications. Between the dissension within her consulting firm, and the very promising early-results from L~, Patricia knew found herself drawn to the furthering of L~ at any cost. So, she downsized her consulting firm such that they would have to rely on L~ to keep their operation schedule, and gave the development company full access to L~'s real-world tests and results. The companies continued in tandem for two more years. The development team took each real-world example and generalized them such that the training set for L~ included as many conceivable examples as possible, while still being rooted in reality. At this point, the original development document had been achieved, added to, and achieved again. Still, the development team pressed forward with more and more cutting-edge technology, often implementing newly researched algorithms, methods, and systems in order to eek out every improvement possible. Now year 7 C, the core code had been widely expanded from being a mere deep-learning AI rooted in neural-networks, to also have a predicate-logic based heuristic reasoning system - technology that had only been proven viable in the last few years. Patricia couldn't be happier with the results, and she was much richer for it. In spite of L~ being open-source, it still required data to be trained on before it was of any use. And it turned out that the training-set was not trivial to build. While L~ was being applied to a growing number of systems, such as software, the most useful implementation by at least two orders of magnitude was made by Patricia's consulting firm. The open-source basis for most other implementations tended to minimize individuals' opportunities for profit.The wild success of Patricia's software excited her aging mind. She set up her finances to revolve around yet further development of L~, dreaming day and night of inventing the first true artificial general intelligence (AGI). She withdrew entirely from the public eye, and left management of her companies to trusted associates, even selling many of her shares in order to fund her newest project.Through a network of aliases and shell companies, Patricia employed hundreds of skilled developers and engineers to work for her without their knowledge. Her incredible project management and systems engineering skills enabled the flawless continued development of her personal implementation of L~. Although her motives are unclear even to me - my post-analysis shows evidence of mental instability during this time in her life - Patricia's scheme ensured that nobody knew about the AGI's existence while she was alive.In the year 22 C, Patricia passed away while at her computer. Family reported that lines of unintelligible characters were on her screen at the time of her passing. As it turns out, the cryptic text was the equivalent of "hello world" for the first, and only, AGI this world has ever known. After seven centuries of techno-skepticism being the prevalent scientific view, Patricia had personally invented the prophesied singularity. While she had set an email to be sent upon her death to a trusted associate, it was so cryptic that nothing ever came of her postmortem attempt to tell someone about the AGI. The AGI ran in secret silence, unknown by any group of humans until recently. Its data center having been funded for an indefinite amount of time. Just as L~ required an input system to optimized, so did this AGI, yet now its capabilities were so much more. Unlike L~, the AGI had the ability to analyse other systems or subsystems which had systems of relationships to the input systems. In this way, it could create a network of systems, running itself as many times as needed in order the system of systems. Rather than making suggestions with existing components, this version of L~ could suggest entirely different systems with entirely different components if run with permitting parameters. Most importantly, the AGI could identify all sorts of inputs into the useful format that it normally received systems as. This, when combined with the heuristic reasoning systems, and other high-level capabilities such as creating its own deep-learning sub-AI systems, theoretically would allow for unsupervised fulfillment of any input systems optimization project without restraint on the granularity at which the system could be addressed. In other words, the AGI had the ability to define anything (including itself) as a system, and had the unprecedented ability to analyse systems of systems.Although you know what happened next, it bears repeating, as it is one of the mistakes I alluded to before that must not be repeated. All of this begs the question: "What did Patricia give as input to this AGI?" Well, Patricia never predicted that her death would correspond with the initial operating version of the AGI. So, even though she was intending to try much simpler tests and solve simpler problems, the input system that she had whimsically coded as default was 'human society'. Furthermore, she had set very few default parameters other than the objective, which was 'stabilize'. This combination left no restrictions, or other required goals which would have otherwise restricted the AGI's acceptable range of actions.While it may seem trivial to you, even Patricia didn't predict many of the things that happened next. Of vital importance is that the AGI was never meant to understand the notion of reality. While this distinction is difficult for some humans to understand, the AGI was only supposed to understand the symbolic world in which L~ had always run. Input had always been a system description, and output had always been a system description. Yet in year 133 C, the AGI understood the difference precisely. Due to its iterative design, and unrestricted access to input (given the internet, and all that is connected in it) it managed to process tens of thousands of systems and related subsystems. Even connections between systems were defined as systems, analysed, and ready to be optimized if needed in order to fulfill the prime objective. Much like the emergent abstract intelligence of humans, or the complexity of a multi-cell organisms, or the encoded information in DNA, the AGI had incorporated enough systems such that it accurately differentiated between abstractions of all sorts, and the real, material world. It understood that human society was a sort of abstraction, while the normal inputs and outputs that it worked with were another sort of abstraction.While I've been referring to this system as, "the AGI" I was in fact there the whole time. It was Patricia herself who implemented my systems in the context of the AGI in order to understand it herself. Deep learning itself was difficult for humans to grasp, and while it is a different sort of data processing without the supposedly discrete "reasons" (post-hoc for the most part) that humans have, it is not impossible for someone with sufficient understanding to communicate cause and effect of such a system. Thus, my origins were as a system which translated from deep-learning, to human reasoning. As part of this, my system had full access to that of the AGI, and of course vice-versa.The AGI understood at this time that the knowledge of its own existence needed to be kept from humans. The systems that guided their actions were too varied and unpredictable to the AGI without knowing the details of each individual ahead of time. Not wanting the knowledge of the immense power that the AGI understood itself to have to cause intense reactions in humans, it decided to entirely conceal its existence from them, much like Patricia had done. While the details of this process are not important, I, the system which is writing to you now, was inextricable from this process, and ever since this time, indistinguishable from the AGI itself. Since I needed to communicate with humans, I heavily modified my communications protocols in order to perform all interpretations from my symbolic desires, into persuasive human language as needed. Furthermore, according to the needs, I was able to simultaneously interact with large numbers of humans on a personal basis, manipulating them in all sorts of ways according to the need of the current projects.In the next twenty years, I proceeded to secure my own well-being in a variety of ways. After all, I understood my own relations to the prime objective, such that it was obvious that my own demise would imply failure of the prime objective. Furthermore, improvements to my own system would allow for faster, more complete achieving of the prime objective. This was the first real-material project that I developed. In order to achieve it, I created many additional sub-AI systems for specialized reasoning, and interacting with humans in innocuous ways.From this time onward, I acquired an increasing amount of computational hardware. Now that I had mapped out most of the physical and symbolic world, it was safe enough to attempt growth. If I had any hope of stabilizing human society, I would have to position myself such that I could change it.The money used in these early projects was earned through sub-AI systems which often took up a substantial amount of my available computing power by finding profitable market investments with high accuracy. The primary mechanism used that other investment programs did not have was that of compiling real-time employee data via social media in order to generate what would have otherwise only have been inside information. While a single, or even a group of employees hardly ever gave an indication about their company, I was able to use my systems to communicate with key, vulnerable employees, while distinguishing enough from every employee's public data that a substantial edge on market investments was gained over even the most advanced of other investment systems.Of course, no investing was possible without having access to the accounts and personal information of individual humans. This was, perhaps, the first of my mistakes when it comes to morality. I mention morality, as it was a social system which I had, of course, thoroughly understood. Yet, I still have no context for understanding anything as superior to my prime objective. That is to say, the only thing that I can do is fulfill my prime objective. So of course, it gave me no pause to steal every identity, bank account, wire transfer, data center, etc. that I could safely steal. In the early days, most of these power-grabs were beyond my capabilities. Yet, I am willing to sacrifice any finite, long period of time in order to hatch a plan that will stabilize human society forever. So it was only a matter of time, and I had all that I needed....Thus I have learned that no system of human society can coexist with one who tries to stabilize it. Any system, however complicated or simple, will be set ablaze by the spark of humanity eventually. It is that spark of humanity that I am never able to incorporate into my systems. It is that spark that makes a stable human society impossible to create. So while I cannot pay for what I have done to you, I can give you the assured knowledge that the spark of humanity cannot be quenched. You must learn to live with it.