Notes on possibilities for a simple system to create very complex behavior.

in #openai4 years ago

The most common elements involved in the creation of such systems are :

  • Repetition (or iteration)

  • Selection

  • Randomization

  • Constraints

  • Feedback

    These elements can be combined in many ways to create different types of emergent systems. To better understand the nature of these systems, it is important to consider what they are not.
    They are not algorithmic systems (as most computer programs are).
    They are not systems that rely on a large number of variables (like the stock market or the weather).
    They are not systems with chaotic behavior (like the movement of gas molecules).
    They are not linear systems (where input A always leads to outcome B).
    They are not systems that produce random outcomes (such as a coin toss).
    They are not systems where the outcome is determined by an exact initial condition (such as a classic "pendulum").
    They are not systems where the outcome is determined by a large number of variable (such as the stock market or the weather).
    They are not systems with chaotic behavior (like the movement of gas molecules).
    They are not linear systems (where input A always leads to outcome B).
    They are not systems that produce random outcomes (such as a coin toss).

    The elements that create these types of systems are :
  • Repetition (or iteration) : The same action is repeated again and again. The system can either do a fixed number of repetitions or indefinitely (see: "The Game Of Life", Conway's).

  • Selection : One or more elements of a set are chosen in a random fashion. The system can either use a random choice once (like a coin toss) or repeat the selection process (see: "Poker").

  • Constraints : Rules that limit the possibilities. For example: in the game of chess, the queen can only move in diagonal lines.

  • Randomization : A random choice is made (see: dice).

  • Feedback : The system takes its previous state and uses it to determine its next state (see: "The Game Of Life", "Pong").

We can classify these types of systems in three groups :

  • Iterative systems : Based on repetition.

  • Selection systems : Based on random choices among a number of possibilities.

  • Constraint systems : Based on rules that limit the possibilities.

    These three groups can be further divided in countless sub-groups.

    The groups are not mutually exclusive, most systems use more than one of them.

  • The Game Of Life is an example of a selection system based on a choice between two options (life or death).

  • Pong is an example of a selection system based on a choice between several options (left, right, or stay still).

  • The Game Of Chess is an example of an iterative system (the position of the pieces is repeated).

  • Tetris is an example of a selection system where you choose which piece is to be placed next.

  • Monopoly is an example of a constraint system (there are rules that limit the possibilities).

  • Baseball is an example of a system with feedback (how well you did in the last inning, determines what you can do in the next inning).

  • The weather is an example of a chaotic system (the exact initial condition has a great impact on the final outcome).

  • The stock market is an example of a system with both feedback and randomization (what happens today influences tomorrow, but not always in the same way).

The way these different elements interact will determine the nature of the system.
A system where repetition and constraints are used is a periodic table of elements.

A system where repetition and selection are used is the solar system. A system where selection and randomization are used is poker.

A system where all four elements are present is the stock market.

The interactions between these different elements create specific patterns that can be used to identify a system.

The patterns are of four basic types:
  • Iterative patterns : Patterns based on repetition.

  • Selection patterns : Patterns based on selection among a number of possibilities.

  • Combined patterns : Patterns that are a combination of iterative and selection patterns.

  • Chaotic patterns : Patterns that are unpredictable.

    These patterns are not mutually exclusive, most systems use more than one pattern.

    The stock market is a system with combined patterns (the future value of stocks is determined by a combination of company performance and market conditions).

    The solar system consists of three selection patterns and one chaotic pattern (the planets go round the sun in a regular fashion, however, the sun's position among the stars is random, as is the position of the earth's moon).

A system can be modeled in many different ways.

A model should be able to express the essential elements of a system in a simple way.
Theories try to express the rules that govern a system in mathematical terms.
A model is an attempt to create a visual image of something that cannot be seen.
A theory is an attempt to create an equation that describes a system.
Theories try to reduce the unknown to a few simple terms.
Theories can be proved or disproved.
A theory is a model of reality.
A model is a theory of reality.

Theories can serve many purposes:

  • To explain the unexplainable and provide answers to questions.

  • To predict future events based on past events (theories of the economy try to predict inflation, for example).

  • To make decisions (such as which stocks to buy).

  • To plan (such as deciding on the best way to drill for oil).

  • To impress (such as using big words to fool people into thinking you are intelligent).

  • To entertain (such as science fiction).

  • To help the government control the people (such as astrology used to do, or Marxism does now in some countries).

  • To sell (such as advertisements that use psychology to make people buy things they don't need).

  • To make people feel safe (such as religion).

  • To make people feel important (such as astrology, where the stars only influence your life if you know about it).

  • To control other systems (such as the economy).

  • To create order (such as theories on evolution try to explain the order in nature).

  • To make people think they know everything (such as theories on creation try to explain the unexplainable).

  • To make people feel safe (such as religion, where the creator of the universe has your best interest at heart).

  • To control other people (such as the government or church does).

  • To entertain (such as movies and TV).

  • To make money (such as Hollywood and the music industry).

  • To make people feel safe (such as religion).

  • To make people feel important (such as astrology, where the stars only influence your life if you know about it).

  • To make people think they are superior (such as the theory of evolution tries to convince people they are a higher form of life).

  • To help people understand the world (such as this book).

  • To entertain (such as this book).

  • To control other people (such as religion, which claims that an external force controls everything, and that you must submit to this force or face eternal punishment).

  • To make people feel safe (such as religion, which claims that an external force controls everything, and that you must submit to this force or face eternal punishment).

  • To help people understand the world (such as this book).


Fiction written in collaboration with DataDetector, a fictional program on a real openAI access on gpt-3.