Human AI Net

Human and Artificial Intelligence Network

(click the picture to see full size)
(or go to http://MouseSearch3d.com to use the one running now. It runs fastest in Google Chrome browser)

Contact:

Ben F Rayfield    benrayfield@mousesearch3d.com    408.334.7214    My resumE (looking for a computer job)

What is it? And what open-source license?

Human AI Net is a free and open-source (GNU GPL 3, or version 2 allowed for most parts) software I'm building that is going to change the world in ways most people can not imagine. The plan is to build a peer-to-peer software to create new kinds of communication between millions of people and computers so they can think more as 1 mind through this system, a mind so far above the IQ scale that its scary, a mind that can be used like a computer, and to use it to do experiments and learn what the laws-of-physics are and generally learn how the universe works and build new kinds of software and games and kinds of communication and other cool stuff. Everyone who uses the Human AI Net is part of the Human AI Net during that time. You will think for it and it will think for you, so everyone can do more than they could before. Eventually I want to use it to figure out the laws-of-physics enough that we learn how to build faster-than-light starships, so we can go look for aliens. Thats just 1 of the many things this could lead to. For now, its a lot of technical plans and a little software, but I've thought about it enough to know how it would work. Its not a question of will it work. Its a question of how long it takes to build it. Human AI Net will be a new kind of life, a new species.

Download:

I've made a little progress, and you can download what I've built so far here:
http://download.com (search for Human AI Net)
http://sourceforge.net/projects/humanainet

How to use it:

Its 1 file. Its a .jar file, which is similar to .exe files. Double-click that file and it works immediately. No setup or installing. You use it through a web-browser. It will tell you what address to type into your web browser after running that file.

Source Code:

For people more skilled at computers: Where is the source-code for you to use to build more programs? All .jar files are also .zip files. Unzip it with Winzip or any unzipping program. It will have a build.xml file. Change the files to build your own programs (under the same open-source license). Use Apache Ant on the unzipped files to create a new .jar file thats similar to the one you unzipped. Thats your new program.

Whats inside the program and how does it work?

Its all in 1 jar file (different jar files depending on which features you want, or all features), but if you want to know what will be in future versions (some of these softwares do not exist yet) of that 1 file that makes it work, read below:

Required parts (in future versions of that 1 jar file):

Javassist - Java compiler inside this Java program, so it can modify itself.

Apache HttpCore - web server, access to web browser and other Human AI Net servers. This is the HTTP type of internet communication. UDP will be done a different way.

Armed Bear Common Lisp - Lisp that runs inside Java. It has a Lisp compiler and interpreter so you can optimize it or not. Human AI Net will be a combination of Java and Lisp, all inside 1 Jar file that you double-click to run. See the /lisp path example in JSelfModify below. The USA Constitution gives me the right to bear arms, and this software also allows me to arm bears.

Evolve Data Struct (my software) - a new kind of database that supports the EXPONENT and PERMUTATION operator. Most databases only support MULTIPLY and PLUS in the ways they do control-flow and combining of queries.

Giga Line Compile (my software) - Compiler optimization for code that artificial intelligence tends to write more often (specificly what Audivolv already writes) which is a lot of permutations and combinations of similar code with variables renamed in some parts of code to overlap variables in other parts of code, and recursions of such permutations and combinations.

JSelfModify (my software) - abstract file system and tree of Java objects. Seamless combination of files in the current Jar/Zip file (the program running now) and files on the hard-drive and things in the program. For example, you can get the file /files/anyfiles/nlmi.html which returns the bytes of that file, and you can also run Lisp code (in the next version I'll upload soon) using paths like this: /lisp/"(mapcar (lambda (x) (* x x)) '(2 3 4))" which returns "(4 9 16)". Whatever you put in the /website path (like /website/files is a pointer to /files) can be used from the internet.

Immutable Sparse Wave Trees (my software) - like a String but for waves. Do substring operations efficiently on waves, and do mutations in log memory and log cpu time, a lower Big-O.

Optional parts / plugins (in an upgraded version of that 1 jar file if you want these features):

Audivolv (Audio Evolve) (my software) - Artificial intelligence and access to mouse movements and realtime generated audio (44100 amplitude per second, per speaker).

Natural Language Mouse Interface (my software) - 3d physics and artificial intelligence system that runs in a web browser, controlled by the mouse.

Schrodingers Network Router (my software) - to test a theory of physics.

Lisputer (my software) - to test a theory of physics. Controls Schrodingers Network Router.

Applet Audio Synthesizer (my software)

Free Speech Just Pay Shipping (my software) - Richard Stallman is famous in free software for many things including the phrase "Free as in freedom, not as in free beer. This is a game about the same idea, but updated for how things work today: "Free as in free speech, not as in buy 1 get 1 free just pay separate shipping." People who live in USA will recognize some of those words from advertisements. In most ways, there is only 1 kind of freedom left, the kind you buy. This will be an unusual kind of game designed to change the world in very specific and indirect ways through psychology of letting the players pay game money (not real money) to censor and oppress eachother, until they can figure out how to solve their common problem inside the game, to gain all of their freedom at the same time, freedom from the actions of the other players. In theory, this should result in a Coherent Extrapolated Volition which is the thing thats supposed to precede creating an accurate Friendly Artificial Intelligence which is a specific area of research about what a theoretical smarter-than-Human artificial intelligence would do.

Or create your own plugins for Human AI Net and expand the system however you want, under the same open-source license.

Plugin to do protein folding using the following possible optimization (to see if it does it faster or slower than existing protein folding algorithms):
Before we create life, we should learn how existing life works, so we know the competition. I think we can do it better. When I have time... I've thought of a possible optimization for protein-folding (its a chemistry and DNA thing) simulations, which can be summarized as: voxel probability fields (of possible atom being at that 3d voxel location or not) of holographic (spheres around each 3d point) bayesian (for each point, the probability of each sphere, distance being a bayesian chance, same bayesian network used for the whole software, reused for every point thats calculated) supervised learning of current atom position mapped to final folded atom position relative to the spheres of bayesian chances, with some acoustic time-curved pattern matching to curve the sphere radiuses (using sound matching algorithms) to adjust for variations in the distances between each pair of atoms in the protein being folded while still using the fuzzy information of the distances relative to eachother. The recursive part is some kind of fast holographic statistical recursion bouncing between the spheres (summing the chances on the surface, but approximating it statistically using less data points) and updating the matrix of 3d voxels representing chances. Theres only 1 3d matrix, and it starts with 0 chance where there are no atoms (or do it at the amino acid level instead at the cost of adding types (which type of amino acid) to the 3d chance field). In a much smaller number of iterations than protein folding normally takes, this should converge to the folded state of almost any protein, using a little of the normal protein folding algorithms when it gets stuck, and using a little genetic algorithms (as is normally done in some protein folding simulations) to optimize parts of my algorithm. Theres a lot of possible variations of this, but I've described a general direction of research that appears to be missing from protein folding research and could possibly simulate it in a smaller Big-O on average. For "dimensionality reduction", for example, you could add 50 bayesian variables to the probability field and sum the surface of each sphere using the normal way of combining bayesian networks, many times, one entry in the sum for each voxel, and use that to get more accuracy, starting with random values (between 0 and 1) for all voxels for all 50 bayesian variables, and converging those voxel-bayesian-networks the same way as the other calculations are done. Written a shorter way: acoustic variable time curved pattern matching bayesian supervised learning of protein folding recursive holographic voxel probability fields. Just one more thing for me to experiment with, in this case using the FOLDING@HOME data... later when I have time. When I finish more of my Human AI Net framework, it should be advanced enough to do protein folding calculations without the core being specificly designed for protein folding, only as a plugin.

Plugin to use the existing internet infrastructure (patterns of electricity in its wires and in the brains of the people who use this software) to use quantum/multiverse physics to find and amplify statistical relationships between the brains of the people who use the software. Some people would call it magic or telepathy, but I call it physics and science and experiments. I know it will work if we fine-tune this software enough because I've done similar things with my brain (the normal way, not with computers) enough to figure out how it works. I've got some theories and am going to test them. Its not enough that I know how it works. I want computers to do it. Then it permanently becomes a branch of science, and we learn more about physics and psychology from it.

That's how you work toward learning how to and finally building and mass-producing these (by the year 2040 I'm estimating):




If you want to fly, help me build Human AI Net and we will use it to figure out the laws-of-physics. We don't know the laws-of-physics well enough yet because we can't predict quantum physics. They call part of it "random" but I call it "go figure it out and stop making excuses like saying it cant be known or is random". If you want to fly between the stars by 2040, help me do this. Once you build the first starship, the rest aren't nearly as hard. We can build enough for all 7 billion people and replace cars. Theres many ways to help. If you don't know how, call or email me and ask. My contact info is at the top of this page.

About the "random" part of quantum physics. They like to say that because they believe in "free will", but there is no contradiction in there simultaneously being "free will" and knowing exactly everything that will happen, because if the Kolmogorov Complexity (the amount of information after compression, like size of a zip file) of the universe is 0 (and I think it is), then the most extreme form of multiverse theory is true (which simply means the universe equals all possibilities you can write in math), and the future (and past and all other directions of time, like "parallel") branches into all possibilities, which are all equally real, and which path you take is "free will" but the combination of them all can be exactly calculated, so there is no contradiction between "free will" and "determinism". Just think of it like calling the universe as a lambda function (like in Lisp or the Lisputer software I'm going to build) that returns an infinite set (called "your current view of the universe") of chances (1 for each possibility), and the lambda function is deterministic but which part of the return value you use (which future you experience) is more like random or "free will". Therefore, if the Kolmogorov Complexity of the universe is 0 (and I think it is), then theres a lot more to learn about quantum physics, and the starships will work when we research, experiment, and then build them. I wrote more about that (and other subjects) here: http://spacecollective.org/Wildcat/2459/Mathematics-itself-is-on-fire