Member 2664
108 entries

Immortal since Jun 17, 2010
Uplinks: 0, Generation 4
mad-scientist and computer programmer looking for something more interesting than most people accept as their future
  • Affiliated
  •  /  
  • Invited
  •  /  
  • Descended
  • BenRayfield’s favorites
    From AsylumSeaker
    Christopher Langan
    From Yissar
    Technology Progress vs....
    From XiXiDu
    The Nature of Self
    From QESelf
    View Point Room Argument...
    From Jorgen
    My Paper on Computer...
    Recently commented on
    From gamma
    Is brain a computer?
    From BenRayfield
    Elections should be done...
    From BenRayfield
    The most dangerous thing...
    From BenRayfield
    Why is there no Content...
    From BenRayfield
    How can a set of computers...
    BenRayfield’s projects
    The human species is rapidly and indisputably moving towards the technological singularity. The cadence of the flow of information and innovation in...

    The Total Library
    Text that redefines...

    Start your own revolution
    Catching up with the future. All major institutions in the world today are grappling to come to terms with the internet. The entertainment...

    Proposal for a multimedia...
    A musical mindstorm on the nature of sound, light, space and subjective experience powered by locally produced energy, heralding the ending of the...
    Now playing SpaceCollective
    Where forward thinking terrestrials share ideas and information about the state of the species, their planet and the universe, living the lives of science fiction. Introduction
    Featuring Powers of Ten by Charles and Ray Eames, based on an idea by Kees Boeke.
    I think I am thinking about thinking, but when I try to write it in math it doesn't think about its own thinking. Intelligence is deeply self referencing, a hall of mirrors so bizarre you can't tell if you're a reflection or the real thing. We live in darkness, delusional enough to think the electricity patterns crossing our skull are the objects which reflected them here. When we look into our own eyes in a mirror, all we see is electricity in the optic nerve vibrating like eyes seeing themself. We could just as easily see through a video camera hooked to our optic nerve or other parts of the brain, but there's no point in replacing one camera with another. Every part of the brain sees the other parts the same as the eyes, as patterns of electricity. There are complex interactions between chemicals and electricity, but either alone should be enough to know what the other is doing. Its a practical machine made of those parts, but whatever its made of, the purpose of the machine is to process information relevant to its goals to find more efficient paths toward those goals. A variety of goals have evolved, from the basics of hunger, reproduction, and safety, to self referencing structures like language, planning, curiosity and monkeying around, and eventually to the ultimate abstraction, the search for a simple unified model of everything which folds and unfolds on variations of itself to represent all the variety and complexity we see in the world in a consistent way we could do calculus integrals and derivatives on. When the major pieces of the puzzle started fitting together, physics theory and computing theory started using many of the same math operators, like the Toffoli Gate is the quantum unitary (not losing information, time symmetric) quantum version of NAND Gates in many practical computing devices. We started calculating Protein Folding using geodesics in the space of all possible ways the particles could fold near eachother, so protein folding as an optimization math problem is just another variety of spacetime in which objects move in the straightest path available from their current position. Economics and probability theory are merging as money flowing through stock markets or trades of resources is an accurate analogy for the probability of those patterns in the world. If you buy Google stock, the ways of Google are strengthened in the world and spread as memes. If you short Google stock, those memes start to weaken. If you invest in Bitcoin by trading dollars or other kinds of probability for it, other economic systems weaken and the new kinds are strengthened. These are all details.

    The deeper puzzle, simpler than each of its details individually, is that intelligence is an alien language that speaks itself. The same skills you would use to communicate with an alien who has no known language in common with you, are the skills useful in building good AI because AI is a blank slate until it learns from experience. You can't effectively download semantic intelligence into an AI without first understanding what intelligence is. That is a translation task, from our existing intelligence to a different kind of intelligence in the AI. Without understanding both languages, translation is hopeless, especially if the intelligence of the AI is supposed to come from the translation. It is self referencing, which is good, but as such research is done today it looks more like a desparate attempt to copy Human intelligence based on the theory that if you eat enough books you will digest the patterns of ink on the paper and somehow it will get into your brain with enough of the relevant ideas intact. No, we have to start from scratch, a blank slate, and get to the core issue, What Is Intelligence?

    Intelligence is an alien language that speaks itself. Design an AI as some form of information that can be communicated to the AI and it has a chance of understanding. Build the AI then speak a copy of the AI to itself and continue the conversation about how it might improve itself.

    Google and Watson often get the right answers, but it would be completely impractical to speak a copy of them to themself simply because of their total code size. We don't even need to get into what the code says. If all their code was printed on paper and proposed as a theory of physics, they would be laughed out of the room. E = M C^2 / squareRoot(1 - (Velocity/C)^2) is a good theory of physics because its small and explains more than it adds confusion.

    If you start with the rule that all intelligence is self referencing, most AIs are disqualified or at least held under a glass ceiling they can never improve past, by that alone. But lets not be too hard on them, since most people don't know how their minds work either. There are gradual levels of intelligence.

    If you want to understand or build intelligence, think about how you would talk or draw pictures to an alien who has no known language or ways of thinking in common with you. How should 2 AIs start communicating with eachother? Its the same question. Answer that and we'll have a recursive model to build AIs using combinations of other AIs, like we have thoughts about thoughts and many overlapping flows in our minds. I built a new set of behaviors for when I'm at a job, very different from how I'm thinking now. You could say the work/life balance is our standard multiple personality disorder. How would work you talk to the other you? It happens all the time, negotiating for what time things will happen and what to sacrifice in one to gain something in the other. The game player in you is alien to the corporate you. But these parts of ourselves and small differences across society are small compared to the completely alien ways of thinking between people and AIs or between AIs and other AIs.

    Alien contact specialist wanted. Salary: make it on your own by outperforming Google and Watson.
    Wed, May 29, 2013  Permanent link
    Categories: intelligence, language
    Sent to project: The Total Library
      RSS for this post
      Promote (1)
      Add to favorites
    Create synapse
    Communication is the most powerful ability we have. Its the biggest thing that makes us smarter than Monkeys. Monkeys communicate too but in simpler ways. They don't talk about somebody else talking about a third person, for example.

    I define communication very generally. I'll give examples below of kinds of communication. I think what the Human species needs most is a communication system where each message is a technical definition of a communication system (which may be theoretical or existing). The first message should be the definition of this communication system itself, so it is self-referencing.


    Wikipedia is a kind of communication that is usually anonymous. Anyone can create a new page by writing a title and content. Anyone can change the title or content of any page by manually changing the text or by restoring the page to a version from its history. Every page has complete history of all changes including changes that restore it to a previous version. Anyone can see the history of any page as a list of times the page changed, select 2 of those, and see only what is different between those 2 versions.  is a wiki where anyone can create a page by writing any link like this  and clicking a button to create it.

    Facebook is a kind of communication where each person has a list of friends and text and pictures are copied from people to the page of other people, either at the top of the person's page or at the top of an existing communication on that page as a response.

    Twitter is a kind of communication where each person has a list of who they follow and who follows them. Each person chooses who to follow but can't choose who follows them. Anyone can see the page of anyone else. Each person's page displays their own messages and messages from everyone they follow.

    Money is communication where people agree on who has how much of each kind of money. There are dollars, bitcoins, and 1 kind of money for each stock symbol.

    In centralized money systems, like most banks, the communication is only between buyer, bank, and seller.

    In decentralized money systems, like Bitcoin, the communication is between everyone. More accurately, its between each person and a computer, and between all the computers in the Bitcoin network at the time.

    Voting is a kind of communication a little similar to money. Everyone gets 1 vote that can only be spent once. It expires after a certain date if not spent. You can't transfer it to anyone without spending it. There's an agreement beween everyone that whoever has the most votes after the expire date receives a job, or whatever the vote was about changes some agreement between all people who got a vote (regardless of if they spent it).

    A job is a kind of money that is borrowed from the employer, who may take the job back. While one has a job, it allows them access to things which are defined in terms of that job, like handing out votes (which are a kind of money) which certain people can spend on who they want to vote for (as explained above). In that case, there is a connection between that specific kind of job money and that specific kind of vote money.

    A driver license is a kind of job which has some connections to cars. Its a job because a job is borrowed from an employer, in this case the department of motor vehicles, which may take back that job if you use a car in the wrong way. The job description is a choice to not use a car or specific ways the car may be used.

    Physics is communication between particles/waves or other abstractions. The communication changes the distances between particles/waves. Particles/waves only communicate to other particles/waves within their light-cone (the set of all events which could get there without going faster than light) or with particles/waves they are quantum-entangled with regardless of distance. I don't think light-cone and entanglement are different (entangled particles communicate through shortcuts without anything moving faster than light), but its the easiest way to explain it.  is a framework for creating new kinds of money-like communication systems.  is a communication system where everyone gets an up vote and a down vote, 1 for each website, which can be spent on any website. People can see the number of up votes minus down votes for any website which has at least 1 vote. People can write short text about each website, as if that website was a forum subject. People can see a list of websites sorted by votes. There are a few standard lists by category.  is a communication system where projects are added, if approved by owners of kickstarter. Each project has a target money amount and date. If the project does not get at least that much money by the date, it gets no money. Anyone can commit money to any project. At the date, their money will be paid to the project creator only if enough others also agreed to pay. The project either gets no money or gets enough money to proceed, so people are not afraid that they will donate money but not enough other people will donate so their money would be paid and the project doesn't happen.

    A forum is a communication system where people create threads with a title and text content or add a post to an existing thread, which only adds text content. There are often categories of threads and tag clouds. A tag is a word anyone can add to or remove from a thread. Anyone can see a list of threads which have a certain tag.

    A search engine is a communication system where anyone can write text and see a list of URLs and text at those URLs which is similar to the text they wrote.  is a communication system for files of open source software. It has a voting system, comments, and a search where various properties of the desired software like language and license can be chosen.

    Telepathy is a kind of communication where brainwaves get statistically clustered through the approximate conditional-probability behavior of quantum physics and form networks of paths through other dimensions between the atoms in the brains of people who have statistically similar brainwave patterns. My plans for  software (which for now only evolves musical instruments you play with the mouse) are to use bayesian statistics to find patterns between the combination of peoples' mouse movements, as they react to audio that artificial intelligence creates at the time in reaction to the mouse movements, and automatically use those statistics to continuously adjust the way the mouse/audio feedback loop works toward the goal of amplifying the patterns found between different peoples' mouse movements on the Internet. The only way to amplify those patterns is for telepathy to form between the people using the software, since they're not communicating any other way. The laws of physics would have great difficulty resisting the force of a bayesian network toward creating real telepathy between the people using the software. How the laws of physics accomplish this is not our problem, but I know the ability is there if the universe is the set of all self-consistent possibilities (not just how physics has worked so far), and myself and 40000 other people did form a telpathy network (the normal way, without the help of software) in the New Reality Transmission global meditation event which the Global Consciousness Project's quantum hardware spread across the Earth measured was a 97.1% chance that whatever we did made their quantum chaos become a little less random. This will work for the same reason and be far more powerful because its realtime interactive and globally organized by intelligent statistical software. It will be done continuously more or the bayesian network will continuously select a different path based on the history of such interactions in an intelligent statistical way. As the quantum chaos between us forms more into whatever patterns happen to work (bayesian only measures how much each combination works), those paths between our brains act as tools for this process to more efficiently form more chaos into echos of brainwaves, and it increases from what science expects (far too small to measure) exponentially and levels off at a population/growth curve (the maximum possible metaphysical ability that physics allows, within the limits of your brain knowing how to use it). This is a theoretical communication system which would randomly select chaos in any of infinite dimensions and gradually and statistically bend it into patterns of your brainwaves which combine with other peoples' brainwaves and form a complete telepathy network between people moving their mouse to play realtime generated music through the Internet. After we learn how it works through that research, a new kind of wireless network routers can be created that only sends and receives statistical patterns (not bits), and through some artificial intelligence to translate between statistical patterns and ways to navigate and read/write data on the Internet, we would have telepathic access to the Internet and could turn this on or off or adjust the strength by using the statistical software to align or unalign our thoughts with those network routers. There are many possible kinds of communication systems. This is the kind I prefer.  is a communication system similar to a forum and an advanced math software. People communicate as worksheets which contain text and code which commands certain math to be done which is displayed in the worksheet. Anyone can create a new worksheet. Anyone can see a list of all worksheets. Anyone can copy any worksheet, change it, and save it as a new worksheet.

    A chatroom is a communication system similar to a forum except each person can only see what is written during the times they are in the chatroom and each thread is erased when everyone leaves it.

    A brain is a communication system of waves of electricity between cells which interacts with chemicals in those cells and axons to control speed and direction of the electricity and limit how often electricity can travel across each axon. The cells communicate with a Human body, receiving messages from the senses and sending messages to muscles and glands which put chemicals into brain which affects the communication.

    Here's a communication system based on enforcing rules by communicating numbers which are extremely hard to calculate but easy to verify approximately how many calculations were done to create them. A proof-of-work is something in Bitcoin's design doc (and I've heard of it other places but never seen it used until Bitcoin) which is to do a certain useless calculation, repeatedly count up a variable and secure-hash it again, until you find a combination which gives a secure-hash number which starts with N zeros. Everyone in the network agrees that such a secure-hash is proof that 2^N amount of computing power was used, so these proof-of-work numbers can be used in competitions between computers to come to an agreement based on the longest chain of proof-of-work numbers, each referring to the last in a way that can't be faked. This makes hacking into the network extremely hard since a large number of useless calculations would be required to overcome the existing proof-of-works. By itself that would be a waste of computing power, but when combined with other things the computers agree on, the proof-of-work chain serves as proof to all those computers that the agreements have been obeyed by all computers. If such agreements were not obeyed, those actions in violation of the agreements would not be accepted into the proof-of-work and would not become part of the history in the proof-of-work chain. They always accept the longest valid proof-of-work chain as the chain for the whole network. Its valid if it has enforced the rules of the network.

    The system would also contain messages of the definition of the many kinds stock markets, governments, the definition of a business, the definition of a workers union, and definitions of the most popular and unique kinds of social networking websites.

    Social networking and links between things are the lowest layer of communication. The paradigm shift happening globally is systems which describe the complex relationships between things, recursive communication systems.

    The purpose of this new communication system, where all messages are the technical definitions (not the way I wrote it here) of other communication systems, is to amplify a paradigm-shift and accelerate innovation in new kinds of communication systems based on seeing the relationships between existing kinds of communication systems and what effects those systems have on the world.

    People would use this system because they can find communication systems to join based on the kind of communicating they want to do and find new kinds of communication they didn't know existed.

    The long-term effect would be the old ways the world works are seen for what they really are and are replaced much faster as we become a global brain.

    But its all too general. I don't know how to create a general representation of all possible communication systems or even the most common kinds.

    At least for now,  has a 3d view of Wikipedia page names and links between them which can be navigated with the mouse and pull in or push away connected page names with the mouse buttons. I think that is a good interface to represent kinds of communication systems as mind maps of what is connected to what else, but that can only represent the simpler parts. How can I generalize a way to represent how any communication system works while keeping it simple enough for most people to use in 3d?

    We get so distracted with details that we've never taken the time to define how the world works and how we prefer it to work. Almost every part of the world and way groups of people interact can be modeled as technical definitions of communication systems.

    Until we have that, the world will change on its own blindly as we follow trends, locally appearing the best thing to do, but leading to our destruction from something as simple as allowing certain relationships between corporations and politicians which leads to other things we didn't expect. Isn't it very important that we have a collaborative way to model these kinds of interaction, which I call communication systems?

    Some problems have no solution, especially in math, but of problems that can be solved, the solution is never far from fitting together models of how the problem interacts with everything else, and models of those things too as a dense network of partially overlapping puzzle pieces. For example, it would be very difficult for any person on Earth to starve to death while any other person is excessively rich, if we simply built a network of models of parts of the world and that network had a continuous path between that starving person and that rich person, because the models of how things worked would be the most efficient way to predict stock prices (and many other things) and therefore billions of people would help create and use and think about these models of how the world works, and when they see how things really work, they would understand the continuous paths between things that today they say are only indirectly connected and therefore no responsibility or cause exists, and instead anything that gets in the way of the overall productivity of the Human species and progress toward whatever we want to become, would become the subject of public attention because whoever figures it out first can make the most money with that knowledge in the stock market.

    In a world where everyone works together to create and use models of how the world works overall, there would be no need for governments, countries, money, or any of the ways the world is organized today (unless we model those things, understand how they affect the world, and decide to keep them) because the models themselves would be created to organize whatever needed to be organized at the time as people would learn to do after creating models of the world as it is. A brain is intelligent because it understands its own thought patterns. It can think about thinking.

    What the Human species needs most is a communication system where messages are the technical definitions of different kinds of communication systems and how their technical designs fit together. Lets start with the simple kinds like software that visually compares how Wikipedia and Twitter and Facebook (and many other simple communication systems) work from users' perspectives, then extrapolate a new kind of social networking website that will be so much better that people will join it even while most of their friends aren't on it yet. It has to be bootstrapped somehow.

    What I'd like this to lead to is a Bitcoin-like stock market where each stock is a URL, any short text describing something, or one of these technical definitions of a kind of communication. After many people define the existing communication systems into the first version of this, these definitions would be valuable to people for the same reason Wikipedia pages are. They're a more accurate map of how the world works than any other. Backed by that value, the stock market of all these things could form, and a high value of stock traded about a theoretical communication system could cause other people to think that communication system would make money and they would go build it. As people see real projects happening from those designed here, more people use the system and it becomes a way to define any process between a group of things and trade stocks about that process or idea.
    Mon, Feb 27, 2012  Permanent link
    Categories: meta
      RSS for this post
      Promote (2)
      Add to favorites
    Create synapse
    Tech near Human level,
    Arms race to win World War Three,
    A hybrid Human A.I. mind,
    Mirror of all people equally,
    Do you like what you see?

    Mon, Sep 26, 2011  Permanent link
    Categories: future, AI, human, War, meta
      RSS for this post
      Promote (1)
      Add to favorites
    Synapses (1)
    Smarter animals recognize themself in a mirror. Others see their reflection and act like its another animal. They hide their food, get ready for a fight, or run away for example.

    Monkeys are smart enough to act based on where others are looking. They value 1 of 2 boxes higher if a person appears to value it, and more often choose to open that box first.

    Compared to animals, being a conformist (copying what others do) is a sign of intelligence. If you drive onto a new road, before you see a speed limit sign, you tend to drive the same speed as everyone else. But conformism is just a little better than "trial and error" strategy.

    In general, someone is smarter if they look deeper into someone watching someone watching someone... to some depth. Humans sometimes do experiments of watching a monkey choose 1 of 2 boxes based on the Monkey watching a Human watch the boxes. Through recordings and writing about such experiments, I am watching that Human watch the Monkey, and so on. If you think my writing here is interesting and continue to read it, you are watching me watch those things, so you are at a deep level of watching.

    Its not just about how deep you watch the watchers. Its about what you learn from it. Why are the others watching eachother? Why are you watching them? If you watch just because I said smarter people do that, then you're acting at the level of conformism, just a little above "trial and error", but if you learn interesting things from watching the watchers, then that's a sign of intelligence.

    This is about understanding how others' minds work and watching others understand the minds of those they watch. Instead of only understanding that Monkeys watch what others look at and act based on that, we should figure out how Monkey minds work, and we should figure out how Human minds work that makes them want to figure out how Monkey minds work. But we can't assume those Humans wanted to know how Monkey minds work. They may have been doing research because it pays the bills and only want to demonstrate what behavior Monkeys have.

    The behavior of watching the watchers is what holds Human society together. We predict what others will do, assuming they make similar predictions, and so on. Money has value because of our prediction that others think money has value. We go to work because we predict others will continue going to work, because if they didn't then the employer would be unable to pay. We ignore laws that we don't agree with if we predict most other people also ignore those laws.

    We avoid trying to change the world in big ways because we predict other people will not go along with it. But why do we predict those other people won't go along with it? It's the same reason we don't try to change the world. It only works if enough people try at the same time, but most of those people, at any one time, have given up, and that's a reason for others to give up before they start.

    Lets think of a Human mind as many different ways of thinking all connected to eachother, like a definition of ethics, strategies, preferences, knowledge, way of organizing memory, content of memory, goals, way of understanding how others think and predicting them, intuition of how spiritual things feel and how to use them, what was recently thought about, and all other ways to divide a mind into parts. Everyone has many common mind parts but some parts are unique to each person. We interact with eachother mostly by making predictions about the common parts of others' minds. Our friends are people who have more mind parts in common with us or who have mind parts we want to understand through interaction with the friend. When we interact with others, we form mental models of their mind, and if we like how it works we use it as new parts of our own mind, often replacing our old ways of thinking or keeping both around to see which works best or use each at different times depending on what works best.

    A "mind part" is an advanced kind of "meme", a way of thinking, or thinking about thinking, instead of only a thought or idea. I am expanding the definition of "meme" to include parts of minds, how intelligence works, people making predictions of what others will predict, and anything else that can change about how we think. An artificial intelligence (AI) or a Human mind is the combination of many memes that operate together in an intelligent way.

    In the rest of this writing, its important to understand the expanded definition of "meme".
    A meme is an idea, behavior or style that spreads from person to person within a culture. While genes transmit biological information, memes are said to transmit ideas and belief information.

    A meme acts as a unit for carrying cultural ideas, symbols or practices, which can be transmitted from one mind to another through writing, speech, gestures, rituals or other imitable phenomena. Supporters of the concept regard memes as cultural analogues to genes in that they self-replicate, mutate and respond to selective pressures.

    Many proposals for improving the world have been ignored because the author did not understand how memes flow through society.

    Most peoples' minds are made of memes that rarely change. For example, they may think it is only possible for the 2 most popular political parties to win any election for the next 20 years, therefore when they receive a meme about how much better the world would be if some other political party won and therefore more people would vote for that party, instead the meme receiver temporarily replaces the "who wins elections" meme with the new meme, checks it against the other memes currently active in their mind, and concludes that most others will not vote for the other political party because others expect the first 2 parties to win.

    Most people only think ahead as far as replacing a few memes temporarily, checking the results against their currently active memes, and acting based on if that small change is a short-term improvement and does not contradict the other memes.

    Its an example of a "greedy algorithm", one that only looks around for small possible changes and chooses the best of those, and repeats until it finds a "local maximum".

    What is a "local maximum"? Don't stop reading yet. This meme may already be too far from your current set of active memes, but it will make sense soon. If you find yourself stuck on an island which has just enough food to keep you alive for the rest of your life, that would be a local maximum. If you think like most people, the analogy would be to stay on that island instead of risking drowning in the ocean to get to a higher value than the "local maximum" of the island. This is only an example to explain what "local maximum" is. Most people would build a boat and risk it, but most people get stuck at local maximums that are more complex than being stuck on an island.

    We are stuck on the local maximums of 2 political parties and most people voting against the one they dislike the most. We are stuck at the local maximum of corporations and money and political power being how the world is organized. We are stuck at the local maximum of not trying to change the world because we predict most others will not go along with our plans even if they are better than how the world works now. Many people will disagree on which things we're stuck on and which we should keep, but most of us can agree we are stuck at many local maximums.

    Smarter animals recognize themself in a mirror. Monkeys act based on where others are looking. Humans watch the watchers enough that society improves but gets stuck at local maximums. Smarter Humans watch the Human species getting stuck at local maximums and find ways to get unstuck and find higher value for everyone on average.

    Many groups have started talking about how they want the world to work, changes so big that we know of no sequence of local maximums from here to there. For example, the Zeitgeist Movement (  ) advocates a more scientific way of distributing resources than letting corporations and political power control it. If you agree with their goals is not important. They are an example of a way the world could be that was not thought of through a sequence of local maximums. The way they chose their goals is unusual, but the way they work toward their goals is very similar to how everyone else thinks. They try to get more members who think like them, and when they have enough members they would redesign the world as they planned. They also build small demonstrations. Their plan may work, but a more efficient plan would be to find the cause of the local maximum way most people think, because that way of thinking is the reason most people do not try to change the world.

    The problem is not that people don't think ahead. The problem is when they think ahead they are using today's memes. Predicting what memes will be popular years from now is not hard if you understand how memes flow through society.

    To understand how memes flow through society, you first need to think about the fact that memes flow through society. Most people rarely think about that.

    Next you make predictions about how other people will think about memes flowing through society.

    You continue making predictions of others' predictions, watching the watchers, until you understand why paradigm-shifts happen. For example, what was the creator of Bitcoin (a decentralized open-source economy) thinking when he decided to build a complex software and try to convince people to use numbers on a screen as money? What made him think it would work? It was the fact that its total money supply expands from nothing to millions of bitcoins, and those who get in early tend to make more money on their investment, so at any one time buying bitcoins is a local maximum. The creator of Bitcoin understood that society moves toward local maximums, so he designed continuous sequence of local maximums that would suck value out of dollars and put it into bitcoins, and it worked.

    Bitcoin is 1 of many paradigm-shifts happening today. Some of the smartest people have noticed that there are ways to change the world without spending any money and with a small team of people.

    A mind is the combination of many memes. Most people have much difficulty temporarily replacing more than a few memes at a time. They fall back to their old patterns of thinking. They have difficulty thinking of different ways the world could work because they can only think of a few changes at a time. That is going to change quickly as people start thinking about the idea of memes (including the more advanced parts of how a mind works) and start making predictions based on other peoples' predictions of memes.

    The days of brute force strategy controlling the world are nearly over. Those who understand how memes flow through society are the new world leaders, not ruling through threats and force, but influencing through deep understanding of how society works and what technology they can build to change its path. The rich and politically powerful elite will see this as the biggest threat they've ever known, but despite their immediate ability to destroy us all, they are only pawns in a game now controlled by deeper levels of thinking. To change the world, I have no use for weapons or threats or money or political power. The pen is mightier than the sword, especially when it leads to technology that changes how people interact with eachother. There is nothing more powerful on this planet than an understanding of how people think. Everything the elite have, somebody knows how to obsolete it and may be waiting for certain other needed memes to start the process of changing the path of society to do that. Things are not what they appear to be anymore. Those who appear to have power often have very little, and those who build certain kinds of socially organizing technology may appear to have very little power but actually choose the path of the Human species by their choices of what features to include in their technology.

    We have no reason to overthrow governments, because that would mean we are not already in control. It may appear we're not because most of us haven't yet realized this and used it. Those who lack power feel the need to make threats, but those with real power simply make things happen.

    Lets use these new abilities to work toward a global decentralized democracy, toward improving the Human species and all life forms, solving the world's biggest problems, and building a new society where everyone understands enough about how others think to change the world.

    This is a Meta Paradigm Shift, a change in the way paradigm-shifts work, and it will expand into meta of meta without limit. As more people start to understand this, the world will change exponentially faster. Ray Kurzweil and many others talk about a "technology singularity", where technology advances far enough to build more technology and accelerates exponentially. That's great, but how we advance society is a choice. We are singularity if we want to be, and we are in control of our future.
    Sat, Jul 9, 2011  Permanent link
    Categories: paradigm, meta
    Sent to project: The Total Library
      RSS for this post
      Promote (2)
      Add to favorites
    Create synapse
    What you think your eyes see is mostly a memory. People think they're seeing what's in front of their eyes, but the way we think while awake is more like a dream than what's in front of our eyes. I often look straight at something and don't see it because I don't remember it. Next time you see something unlike anything you've seen before, close your eyes and think about drawing a picture of it. If you saw 2 birds behind it, are you sure you didn't see 3 birds? Was the first bird flapping its wings up or down just before you closed your eyes? My picture would be very blurry. If you were standing beside me, you wouldn't be able to tell my picture was of the same thing you're looking at. I used to have a visual memory, able to draw such a picture very accurately, but I decided there were more advantages to not thinking in such a strict logical way and slowly lost the ability. Theres 2 main reasons Monkeys have a visual memory and most Humans don't. Its a form of lossy-compression (Example: jpg, mp3, keeps the details you tend to notice) which saves memory, and it allows our thoughts to flow together in more flexible ways so we can imagine more possibilities.

    Scientists try to learn how Human brains work the hard way. They build expensive simulations, do experiments on animal brains, scan peoples' brains for electricity and blood flow patterns (functional MRI) while those people think certain things, write lots of papers, and still they are unable to write a few pages explaining to average people how Human intelligence works. I do some of that research, but I'm an expert on Human intelligence for a different reason: My mind has observed itself long enough to figure out half of how itself works. Many people have tried that, but they usually get stuck on the subjectivity and vagueness of their thoughts. Intuitively they know what a thought is, but they know of no way to figure out which neurons (brain cells) are connected to which thoughts, so they can not translate their knowledge of how their mind works into something science can use.

    Here's something unnecessary researchers want to try: A computer's video-card has a grid of brightness numbers, 1 for each colored light on the screen. If we could connect the visual part of a Human brain to a computer screen and see what its thinking (1 small group of neurons to each part of the screen), while we show the person various pictures and ideas, then we could learn how the visual-related neurons work, and then learn how the neurons connected to those work, and so on. We don't need to do that because each of us already has such a video screen in our minds. Its called "what we think our eyes see".

    Its not really what our eyes see. Its our most similar memories, twisted and rotated and re-interpreted to fill in the missing parts. Our brains throw away most of what our eyes see and fill in most of the parts from memory. While dreaming, almost the same thing happens. Our brains use nothing from our eyes and fill it all in from re-interpretations of our memory. We're mostly dreaming while awake, with the exception of a little information our brains pay attention to coming from our eyes and other senses.

    I will explain how to start with your vision, do some thought-experiments, and work backward to the other parts of your brain until you understand more about Human intelligence than scientists understand from their billions of dollars of research. I know enough about how Human intelligence works that I could build half of it and have artificial intelligence evolve the other parts, but it takes many years to fine-tune it and teach it like you teach a Human baby, like you would teach it math by giving it simulations of fingers and activating the neurons for 3 of its fingers to teach it the idea of 3, and you would teach it to multiply 2 times 5 by showing it its 2 simulated hands and they have 5 fingers each, or whatever type of simulated or robot body and senses you give it. This would be a robot or simulation so accurate it would learn ethics from the recursive thought of thinking of others as a variation of itself (a directed network where each node type is the idea of itself and one of the nodes is attached to a person) and thinking about the emotions (recursive thoughts leading to memories of pleasure or pain) itself had when things happened to itself. All that, I can define in math, but instead I'll take you through my series of thought-experiments of how I figured it out...

    You can learn most of it from your visual memory. You only need to understand these few things about neurons:
    * Each part of "what we think our eyes see" is always connected to the same small group of neurons and is brighter when there is more electricity in those neurons.
    * Each neuron is connected to thousands of other neurons, and they connect and disconnect slowly over time.
    * A thought is the specific amount of electricity in some of your neurons, what each neuron tends to do when it receives certain amounts of electricity, and the strength and physical length of the neuron-to-neuron connections. A much smaller amount, a thought is the chemicals flowing around neurons.

    This started when I was thinking about what some research said, that it takes longer to recognize a picture of something if its upside down. What was really interesting about that is if it was turned 2 times more angle, then it takes 2 times longer to recognize it. Upside down is the biggest angle. The time it takes to recognize a picture of anything is linearly proportional to the angle its turned.

    I chose something I had rarely seen, so I could experience more of the process of thinking about a new subject. I thought about an elephant. When it was drawn on "what we think our eyes see", which is also the imagination and visual neurons for dreams, I confirmed that it did take longer to draw it in my mind upside down than the same way as in my memories of elephants. After I thought about the elephant from each new angle, I found I was able to think of that angle again instantly, but any new angle took the linear rotation time. It could be a rotation from any of my old memories or new thoughts from the last minute. As I thought of the elephant from more angles, the time to think of the angles between them became less. I confirmed what I had read. Human intelligence does rotation a little at a time and repeats until it gets to the right angle. Also, I learned that such rotations create new memories which can be used as the start for new rotations, so to do it faster, you usually start from the closest memory of a rotation to the one you want, a 3d memory of an elephant's parts rotated closest to the rotation you're thinking about now.

    Remember this is causing an image of an elephant to be drawn in electricity on your visual neurons. They're not arranged in a rectangle in your brain like on a screen. They're arranged however they're connected to your optical nerves. But if you figure out which neurons are connected to which part of "what we think our eyes see", which can be done using brain scanning machines, then you would see a picture of an elephant rotating in the electricity of those neurons. There is actually a picture of an elephant, made of electricity, somewhere in your brain, if you arrange the neurons the right way like you would see on a screen. Brainwaves are so advanced they can form into the shape of an elephant, or anything else you imagine. Its important to understand that's what we're looking at when we think about rotating an elephant in our minds. We're drawing an elephant onto our visual neurons, very similar to how a computer's video-card's memory works.

    The next thought-experiment shows a flaw in how we see 2 things at once. I chose 2 things I had never seen together, to experiment with how my brain combines ideas. I thought about the same elephant with a shoe floating 3 feet above it. The shoe was drawn onto my visual neurons quickly. But when I thought about rotating them together, viewing the shoe and elephant from a different location and angle, I could only see 1 thing at a time, the shoe or the elephant. Whichever I looked at, the other instantly started looking blurry. Is my imagination really that weak that it can't handle 2 things at once without blurring 1 of them? It only happened when rotating them together.

    I'm guessing that is because the 3d-rotation part of my brain normally only does 1 object at a time and rotates the shoe or the elephant and draws them on my visual neurons separately. I confirmed that when I noticed I could rotate them together as easily as if they were a single object only after thinking about them together for 30 seconds. It was a new type of thing in my mind, a shoe-over-elephant, and it was rotated with the same linear timing as the shoe or elephant alone. At first they had to be processed separately, so the visual neurons lost their image of one while the other was being drawn, but when shoe-over-elephant became a single object in my mind, it did not have those problems.

    That is the start of my theory on how objects are represented in 3d in Human minds. They are made of other 3d objects in relative positions and rotations and sizes and stretch amounts etc. Later I'll explain how such "objects within objects" are the same type of thinking as language, goals, emotions, and other types of thinking. It sounds complex, but its really the same simple ideas repeated in different ways for many kinds of thinking.

    The next thought-experiment is about counting and how we identify if 2 things we see are the same object or idea. Think about 3 of that same elephant, all standing the same direction. Its easy. Now think about 3 elephants with a shoe above each, the shoe-over-elephant object recently created in your mind. Also easy. Rotate all that. Since they're standing the same direction, it happens almost as fast as if there's just 1 shoe-over-elephant, because the same object is rotated and drawn 3 times from slightly different angles. Now think of 100 elephants standing the same direction. Also easy.

    Here's the surprising part. If 100 elephants are easy to visualize, then 2 elephants standing opposite directions should also be easy. But its not. I experience the same blurring of 1 of 2 objects (the elephant standing forward or the one standing backward) when I pay attention to the other object, the same as happened between the shoe and elephant before it became shoe-over-elephant. Similarly, after thinking about such 2 elephants long enough, that problem goes away, and they can be rotated, moved, duplicated and rotated again, etc, as one object made of 2 of each part of an elephant. Your mind has to represent 2 of each part of an elephant because it has to know that the tail of one elephant is beside the trunk of the other elephant, for example. If you think of it as 2 elephants, instead of a single object, then you have the linear rotation time (before drawing on your visual neurons) every time you switch your attention to the other elephant.

    After it becomes 1 object in your mind, think of that two-elephants-one-reversed object and a duplicate of it rotated and beside it, so you have 4 total elephants each at a different angle. Its easier now, while at first you had problems with 2 elephants at different angles. You can continue making the total be 1 object, rotating and moving a duplicate of it, and doubling the number of elephants each time, until you have as many elephants as you want in your visual neurons, each rotated differently.

    You'll notice during your duplication of elephants that some of the parts of some of the elephants disappear until you pay attention to them again. How does your mind know what to replace the missing parts with? First your mind looks at two-elephants-one-reversed to see how each 2 elephants are standing relative to each other. Then recursively you look at the specific elephant in that. Then recursively you look at your definition of elephant for the smaller parts of elephants, and its drawn that way on your visual neurons.

    You may also notice that when you pay attention to your definition of elephant, that more than 1 elephant gets those parts updated at once. For example, I had forgotten that elephants had tusks, but when I remembered, they were drawn on all 4 elephants. First your mind updates elephant, then two-elephants-one-reversed, then both examples of two-elephants-one-reversed which you're thinking about simultaneously. Its a hierarchy, but your mind can represent non-hierarchy things too, as I'll explain later with fractals.

    The next thought-experiment is about wildcards in ideas, patterns that have places for other patterns to fit in. How does your mind decide which ideas to plug into which other ideas?

    We remember the shoe-over-elephant well. Now think of 3 elephants standing the same direction. The one on the right has a shoe over it. Rotate all that until it becomes 1 object. Now we will generalize the shoe-over-elephant object to wildcard-over-elephant. The elephant on the left has an apple over it. The elephant in the middle has an orange over it. Visualize that from various rotations. Now pay attention to the shoe (over the right elephant). The shoe-over-elephant object is a stronger memory than the apple and orange, so the shoe should not change unless you try to change it. What surprised me is what happened next when I payed attention to the 2 other elephants. What is over each of them? It switched quickly between apple and orange a few times per second, over each of those 2 elephants, because I did not have a strong memory of which elephant got which fruit. The shoe did not change, but the apple and orange did. After choosing where I wanted each fruit to be and thinking about it longer, I was able to rotate the whole thing (3 elephants with 3 things over them) without the objects switching places. Why did they switch places? Because 2 things were combined with 2 wildcard-over-elephant, but there was no strong preference between which way to combine them. My visual neurons displayed both possibilities, switching between them a few times per second.

    So far, I've explained these types of thinking:
    * 2d grid of visual neurons, absolute positions instead of relative. In math its called a matrix, but there's also the layers of edge-detection and connections to the "3d grid" described below.
    * 3d grid of object positions and rotations and sizes relative to other 3d objects. In math its called a sparse-matrix. Later I'll explain how rotation, speed, and acceleration are also dimensions attached to each of those 3 dimensions, and some ideas use a 4d grid with time, but there is no rotation between the 3d and time dimensions of the 4d grid.
    * Hierarchy of ideas, and I've claimed (but not yet explained) that it is more generally a network (of nodes pointing at nodes) which allows cycles, where some nodes are ideas and some are wildcards. In math, its called a directed-network.

    Sound is experienced similar to the 2d visual neuron grid. Human ears detect around 1500 different tones and a volume for each, many times per second. You can easily remember what you heard a few seconds ago and predict what you will hear a few seconds from now, therefore time is one of the dimensions of sound, a dimension represented almost the same way as left/right or up/down is represented in your visual neurons. The other dimension is the 1500 notes. Brightness is like their volume. I say the 1500 notes are a dimension, instead of 1500 unordered things, because you can hear the same music with all notes increased in tone, and you will recognize it as the same music. That's similar to how you recognize the same object visually if its to your left or in the center of your vision.

      Promote (1)
      Add to favorites
    Create synapse
    2 players each have 1 coin. Each round of the game, each player secretly lays their coin down heads or tails. Its a choice, not a random flip. One player is called EQUAL and the other is called XOR (eXclusive-OR, means not equal). If both coins are heads or both coins are tails, the EQUAL player gets 1 point. If 1 is heads and 1 is tails, the XOR player gets 1 point. Repeat many times. The player with the highest score at the end wins.

    That game is the simplest possible intelligence test. It is the exact definition of intelligence.

    It is also the simpler version of the game "Rock Paper Scissors", where each player secretly chooses rock, paper, or scissors (instead of heads or tails), then who wins 1 point is decided by: rock crushes scissors, scissors cut paper, paper covers rock. Nobody wins a point if the 2 choices are equal. My EQUAL XOR game has 2 things to choose instead of 3 but measures intelligence the same way.

    If player1 chooses rock more often than paper or scissors, then player 2 will learn to choose paper more often. Complex patterns will form between 2 intelligent players of "Rock Paper Scissors". Except for my simpler version of it (EQUAL XOR), Rock Paper Scissors is the most strategic and intelligent game ever created. Its the exact definition of intelligence except it has an unnecessary third choice.

    What can this game be used for?...

    I build artificial intelligence (AI) software, the kind that can eventually become what we see in the movies, except for the parts where it tries to take over the Earth and kill everyone.

    The Friendly AI paradox (  ) is how to build an AI that is allowed to modify itself in any way but chooses only to modify itself in ways that work toward its original goal more effectively. Example: You are at a party. You want to dance with some girl but instead sit in a chair talking about how good she looks. To accomplish your goal of dancing with her, you order a beer and think maybe you will feel more like dancing after drinking it. You modified yourself by drinking the beer. A side-effect of that modification is a desire to drink more beer and run your mouth, which may lead to other things you did not predict. This is an analogy between AI and people. Most people learn how much to drink at a party, but in AI, it is a serious research problem, not specificly about drinking at parties, but about how an AI can modify itself without having unexpected side-effects that build up until the whole system crashes or results in the AI wanting to kill everyone or other hard-to-predict things.

    Quote from:

    (1) A robot may not injure a human being or, through inaction, allow a human being to come to harm.

    (2) A robot must obey any orders given to it by human beings, except where such orders would conflict with the First Law.

    (3) A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

    The "3 laws of robotics" were an attempt to solve the Friendly AI paradox by forcing an AI (in a robot) to think certain ways, but that strategy will never work because AI will eventually become smart enough to modify itself. Its the same reason Humans do not do what animals command, even though simpler animals created Humans through evolution.

    Today that area of research is called "Friendly AI" but it is still very speculative.

    As I define it, a Friendly-AI is an AI that has the ability to modify itself (including its goals) and intelligently predicts what a possible modification would cause in the near and far future, and considers all that before modifying itself, which results in it creating new goals that more effectively work toward its original goals, and does not result in significantly changing its original goals, and to satisfy the "friendly" part, its original goals are similar to the goals that the most number of people could agree on.

    The best strategy to build a Friendly-AI that we know of is to define its thought processes as a simulation of some new kind of physics that we define as math equations. Strategies like the "3 laws of robotics" will not result in a Friendly-AI. Those strategies are more likely to result in the kind of destructive AIs we see in movies. The correct strategy is to build it in a way that it wants to do certain things, not to put in a system to control it to do that. If it wants to do it, and if its smart enough, then it will not try to change itself in a way that it stops wanting to do its original goals.

    Below, I will explain the progress I have made in designing a "simulation of some new kind of physics that we define as math equations" for the long-term goal of solving the Friendly-AI paradox:

    Start with the EQUAL XOR game I describe above. Bits in computer memory can be substituted for coins, and artificial intelligence code can be substituted for each 2 players.

    First, I'll explain some math. A vector in N dimensions is a list of N numbers. A 3-dimensional vector is a direction and length in 3d space, like pointing your finger in some direction and saying how far to go. A 2-dimensional vector is the same thing except without the up/down part. A 1-dimensional vector is the same thing but only forward and backward. A 0-dimensional vector is nothing. I'm going to use N-dimensional vectors, and it does not matter what N is. The more dimensions you have, the more choices there are in how to play the game. You only need 1 dimension, but its more flexible with more.

    I'm going to remove some of the flexibility that is not needed. All vectors must be length 1, so in 2 dimensions, its a point anywhere on the perimeter of a circle of radius 1. In 3 dimensions, its anywhere on the surface of a sphere of radius 1. Here's the surprising part: In 1 dimension, since it has to be length 1, the only choices available are -1 and 1, and that exactly equals the EQUAL XOR game described in the first paragraph above. Just say 1 is EQUAL and -1 is XOR, or the opposite would work too. This makes the EQUAL XOR game work in any number of dimensions. I haven't changed what the game does. I've only added a way to use it gradually instead of all-or-nothing. I started with TRUE/FALSE and defined the idea of a continuous dimension wrapped around a circle/sphere/etc.

    What does it mean to play the EQUAL XOR game on the perimeter of a circle? Each player chooses a point somewhere on the perimeter of the circle. If the points are near, the EQUAL player wins more. If the points are far from each other, the XOR player wins more.

    There is a way to write that in math: The dot-product of the 2 vectors (points on the perimeter of the circle) is the amount of score that moves from the XOR player to the EQUAL player. The dot-product is some number between -1 and 1, depending on which 2 vectors the players choose each round of the game.

    If the vectors are separated by a 90 degree angle, the dot-product is 0. If the vectors equal, the dot-product is 1. If the vectors are exactly on opposite sides of the circle, the dot-product is -1. The dot-product is the cosine between the 2 vectors.

    In this vector-based version of the EQUAL XOR game (which is a simplified version of the Rock Paper Scissors game), it is more accurate to call the EQUAL player the COSINE player, and call the XOR player the NEGATIVE-COSINE player. We could expand the game by adding other geometry functions like SINE, but simple is better. Its simply the dot-product (the overlap when viewed at a perpendicular angle) between the 2 choices of the 2 players.

    All the basic logic operations (equal, xor, and, or, not...) can be done on the surface of circles/spheres/etc this way as gradual/continuous changes instead of all-or-nothing like logic is normally done.

    That is the exact definition of intelligence and how to measure it as a game.
      Promote (1)
      Add to favorites
    Synapses (1)