Global Telepathy Network
Sat, Apr 23, 2011
I offer all software designs in this writing under GNU GPL open-source license version 2 or higher.
I define "artificial parapsychology" as the combination of artificial intelligence (AI), neuroscience, chaos theory, and parapsychology (
), as a bridge between software and whatever metaphysical things brains are connected to, or any machines similar enough to the relevant parts of brains (
) to access the same metaphysical things. Artificial parapsychology will be a new branch of science. Its parts have been researched, but I'm not aware of anyone who has put them together. The first and smallest goal of this "artificial parapsychology" research is to create telepathy between a few people and then a "global telepathy network", a psychic communication between billions of people simultaneously which gets stronger over time, and after that is working expand it to more advanced abilities like telekinesis (moving things with the mind, which I have done many times), starting from statistical psychology software to organize our subconscious thoughts through the internet. Artificial parapsychology is the next step of evolution of the Human species and technology. It could become a way to organize society, a power source, a way to create nanotechnology (the same way some psychic abilities can heal at the atomic level), and a way to rotate out the current statistical patterns called laws-of-physics and use the "Ultimate Ensemble" type of Multiverse (
) to rotate in a different laws-of-physics locally around a power source, like to create a warp field as a more advanced kind of telekinesis done by machines. Artificial parapsychology is the specific way Humans do psychic things, how machines could do those things, how Humans and machines could do them together organized through software and the Internet, and the many undiscovered parts of physics such research will lead to after we understand how psychic things work. Artificial parapsychology is the most efficient way I know to teach billions of people how to do telepathy, telekinesis, thinking things to affect events in subtle ways, and the other bizarre things I've learned to do sometimes (but not any time I want). Anyone can learn to do it, but most never try. Through artificial parapsychology, the learning will be much faster. Telekinesis to small objects sounds interesting but not useful, but it would be much stronger if billions of people, through the exponential chaos-theory skills brains normally use, and the mass of planets and stars etc, were not holding this reality the way it is, strengthing the laws-of-physics and the same patterns of reality that normally happen. If we automate parapsychology and get enough people using the system at once, the opposite exponential effect could happen, to reverse the mental force we put on this reality to keep ourselves trapped in this limited form of existence, and gradually start to have more advanced psychic abilities. In the universe overall, it is unusual to be in only 1 place, only 1 form of existence, at a time. A body is like a tool (evolved as Darwin said, but still usable as a tool, and its a 2-way communication), not somewhere you're supposed to be trapped. We will learn to use many tools at once, many kinds of tools, starting with artificial parapsychology. We can float around in a form more like consciousness. Its a skill most of us have forgotten as we became distracted with life on Earth, but artificial parapsychology is a tool we can use to re-learn it, and to learn many other things.
How will people use it?
In this writing, I will explain how I think these areas of science can be combined and automated with software, to create a new kind of mind made of many computers and Human brains networked together through the internet, using the user-interfaces of realtime generated audio, mouse movements, mind-reading game controllers based on electricity on your head (Like Emotiv Epoc or OpenEEG), and/or any other devices connected to a normal computer. Many user-interfaces will be allowed, and they each are a list of floating-point numbers that change many times per second, like mouse movements are X and Y numbers that change whenever you move the mouse, or a mind-reading game controller is a list of electricity amounts on the skin of your head or a list of numbers derived from that in a software designed to interpret such electricity. Outputs include speakers which are a list of numbers oscillating between -1 and 1 representing each speaker amplitude 44100 times per second for 44.1 khz audio. Input and output are always vectors that change many times per second. That is how people interact with this system.
Find your own evidence:
The evidence for AI, neuroscience, and chaos theory is well known. The evidence for parapsychology is highly disputed, and the scope of this writing does not include evidence for parapsychology, because most people already have an opinion on it and are unlikely to change that opinion, so instead I am writing this only for the people who were able to find their own evidence for parapsychology. Example: If you've ever, for no logical reason, known who was calling just before you picked up the phone, then that's your evidence for parapsychology. Example: The
research about the connection between major world events and quantum number generators. Example: Some rare people, including myself, sometimes (but not whenever we want to) have the ability to move small things (like aluminum foil balanced on a needle) simply by thinking it, which is called telekinesis. Theres a video on youtube called "psi wheel in a box 2" of me doing that. Example: Most religious people think parapsychology exists sometimes or at least in some rare people described in their ancient books. I don't expect you to take these things as evidence, only to look for similar things in your experience so you know parapsychology is real.
The rest of this writing will assume the reader thinks parapsychology is real, and of course AI, neuroscience, and chaos-theory have been proven. Based on that, the next step is to figure out why these things work and combine them.
The technical details of how to build an artificial parapsychology software and/or hardware are almost the same as the technical details of my plans for my Audivolv and Human AI Net softwares for how they will evolve intelligent musical instruments, interactions of moving words on the screen through neural-network-like structures of text, and other things AI would normally be used for. Mostly it will be an interaction of realtime generated music with mouse movements, as the user-interface. The core of that planned design is to evolve data-structures as arrays inside arrays inside arrays... with cycles representing networks of nodes with cycles, and to define the sizes of such arrays in terms of sizes of other arrays in the same node in terms of EQUAL (Example: neural node has same number of child nodes and weights), MULTIPLY, POWER (Example: bayesian node has exponentially more weights than childs), CONSTANTRANGE, or recursions of such array alignments as bigger arrays, and to evolve strongly-typed code that uses those evolved data-structures, and to evolve statistically-typed code and data on top of the strongly-typed system. Artificial parapsychology has never been done in a way as interactive as I'm planning or with this much direct connection to artificial intelligence and is very speculative, but I have my reasons for thinking it will work, and it will be built on a well defined AI system.
The "list of floating-point numbers that change many times per second" for each person can be used as chances in a bayesian network. Bayesian networks are designed to find what causes what else when given many examples of chances for different things, but they do not understand the "different things" as more than variable names. Build a software to find small patterns between these numbers between many people using the software through the internet, and adjust the bayesian network to amplify those patterns. The people will have no direct interaction with other people. They will simply notice the realtime interactive music changing when they do the same mouse movements which will cause them to move the mouse a little differently. The mouse movements are 2 of the bayesian chance numbers. That forms feedback loops between many people across the internet, through them hearing the realtime interactive music and their mouse movements. If this generates artificial parapsychology, then the people will start flowing their thoughts together, at first just through the internet through this subtle music/mouse psychology software, and later through the way psychic abilities normally work between people. The artificial part means artificial intelligence, not just people, will have access to these psychic communications between people.
The "list of floating-point numbers that change many times per second" for each person can be used as a probability balancing feedback-loop between 1 person and 1 computer, between their mouse movements and the AI's response of realtime interactive music. The goal is to put the person into a very balanced state of mind, so all possible thoughts are equally probable to happen next. Max Tegmark, writer of the specific Multiverse theory I think is true (Ultimate Ensemble), also writes that quantum superposition does not last nearly long enough for a superpositioned electron wave to travel between 2 neurons (brain cells), but he wrote that about how brains normally think, not in a very balanced state of mind where uncertainty of which thought will come next is maximized. Using AI, recursively balance the chance of the next thoughts, and the next thoughts after that, and so on, but only statistically instead of by knowing what the thoughts mean. It can be done statistically by using each person's "list of floating-point numbers that change many times per second" as numbers to be influenced toward center position through the subtle psychology of interactive music (to the person) interacting with that list of numbers (from the person).
Schrodingers Network Router is a software I'm planning which will do any of the other technical designs but over the internet using the timing of network communications as some of the "list of floating-point numbers that change many times per second". Quantum physics says UDP packets (the kind that depends the least on other simultanous network communications) going across a network wire take much too long to stay in superposition and use the Schrodingers Cat effect, but that does not mean it can not become superpositioned again when it stops being observed in the wire. In all such quantum tests, it was observed around the same time. There has never been a large scale test of quantum physics to see if it can recursively branch and re-branch over a large area, to see if you branch it enough times you eventually get both versions of Schrodingers Cat interfering with each other in subtle ways through the multiverse. This is about patterns of information through the internet instead of a cat, but the idea is the same.
At each node in a network, run an AI. Each node receives a 1 or a 0 in each communication from other nodes. It uses AI to choose to invert the bit or not. It uses AI to choose which node, of its few nodes that do not change often, to send the next communication to. It may also send some evolving code with the bit, to be combined with whatever evolving code the next nodes have. Such evolving code may modify the way the AI chooses to invert bits or not and chooses nodes to send to next. Each node has a goal ratio of 1s to 0s which its AI tries to accomplish by whatever choices it gets to make (described above). By sending data that changes such goal ratios, through the network recursively, the multiverse can be blurred in some places, sharpened in other places, stretched and tightened, influenced toward statistically selected possibilities.
Chaos-theory through LCD screens, tesla coils in those glass spheres, and other devices, interacting indirectly with the chaos-theory of brains, while people looking at such devices and centering their thoughts causes patterns in the electricity of such devices, patterns which are so small they would not normally be noticed, but artificial intelligence will find such small patterns and amplify them.
In the "How Human Intelligence Works" text file, Human intelligence is explained as using sparse grids made of many example experiences, aligning onto the dimensions of the grids recursively into more grids and dimensions, and directed-networks where nodes are ideas or wildcards, and a 2 dimensional visual grid, and a 2 dimensional (tone and time) sound grid, and a 3 dimensional plus 1 dimension (for time) main grid, and operating from the root goal of recursively avoiding memories of pain and working toward memories of pleasure. Human minds normally think in terms of fractals, in the way that a directed-network with cycles is a fractal, and a directed-network without cycles is a hierarchy/tree. That is half of Human intelligence. If the other half of Human intelligence is evolved (using my Human AI Net and Audivolv softwares, for example) as variations of code where data is a complex type of neural network (N numbers per node and M numbers per edge, each modifiable by evolved code), then through the "list of floating-point numbers that change many times per second", many Human minds could be networked together through the internet in a way similar to how they already think.
In my experience, a few times in the year 2002, small parts of reality would change, like a certain book would stop having page numbers on the left pages (but still on the right pages), and later that same book would have page numbers, which I accidentally caused by trying the "assemblage point" thing described in "Carlos Castaneda: The Art Of Dreaming" (which was the book that changed page numbers). I think that works the same way Schrodingers Cat works, and that its 2 variations of the same book in 2 parallel realities which flow continuously together but people normally don't notice it because they doubt their memory when their memory finds such a contradiction. I've done too many bizarre psychic things to be able to doubt that so easily, when it happened a few times with different objects. The point is parallel realities are real, like in "Ultimate Ensemble" Multiverse theory, and there must be a way to interact with them. In the context of the "How Human Intelligence Works" text file, there must be a way Human brains can figure out if they moved to a slightly different parallel reality, because if they couldn't figure it out then their memory would be changed the same way the past was changed where the book was printed without page numbers on the left. But since my brain could see those few small changes, there must be something about brains which allows them to notice such things. Brains work mostly the same way in all parts of the brain, so whatever works for the whole brain will have much similarity to smaller parts of the same brain. That means we're looking for a fractal behavior of brains, something that works on small and big scales the same way. Therefore we're looking for a way a group of neurons can statistically detect if it has moved to a parallel reality in a parallel brain (also your brain, a different version of you in the same place) so its environment (other neurons) would appear to act differently than they had in the memory of the "group of neurons". If "Ultimate Ensemble" Multiverse theory is true, then everything could move between slight variations of reality (parallel realities) all the time in continuous motions we usually wouldn't notice. My plan is to evolve a fractal neural-network-like algorithm, to be a user-interface between many people through the internet in the Human AI Net and Audivolv softwares, to give our subconscious minds access to such small movements through the multiverse, to observe and influence the multiverse by what we think and how we use the artificial parapsychology software.
How to combine all that:
It sounds complex, but its all 1 system. You can download the early versions and design documents of various modules at
It will be a jar file that you double-click to run and it just works,
a Java program containing a realtime Java compiler for adding new parts to itself,
with API for permutations of function objects and their parts,
where functions apply to simple data structures made of arrays containing arrays containing arrays... with cycles,
which can be used to evolve variations of neural nets, bayesian nets, or any type of network data structure and evolved code that uses it.
It will be a very general system for evolving data structures and evolving code that operates on them.
One of the simpler uses of it would be to evolve variations of a neural net by putting N numbers in each node and M numbers
in each edge between nodes, and evolving a function that takes N+M+N floating point parameters (and weighted-average modifies them) and runs on each pair of connected nodes once per run of the evolved network. Actual evolved data structures will be much more complex.
I don't have to hard-code most of the AI algorithms into it. I'll just define what behaviors I want the AI to have and evolve the AI gradually toward that. I'll write some of what I wrote in the "How Human Intelligence Works" text file and the designs above, in math, and have it evolve some code and data structures to do that. Thousands of people download the jar file, double-click it, move their mouse to play intelligent interactive music, click the "Enable Artificial Parapsychology (possible telepathy, telekinesis, etc)" checkbox, and the "global telepathy network" boots up automatically. Then we don't need the software to keep the "global telepathy network" running, but we can continue to use the software to connect it to the Internet. That, or something like it, is the next step in Human evolution, and its free and open-source (GNU GPL 2+) for everyone, but I could use some help designing and building it.