23 thoughts on “Conway’s Life implemented in Conway’s Life”

  1. It’s implemented using the Hashlife algorithm, which techies may find interesting:

    HashLife is an algorithm created by Bill Gosper in 1984 for simulating the Game of Life. It is designed to take advantage of the considerable amount of repetitive behaviour in many large patterns of interest.

    Roughly speaking, the idea of the algorithm is to store subpatterns in a hash table so that the results of their evolution don’t have to be recomputed if they arise again at another place or time: 2^N+1×2^N+1 tiles are run 2^N-1 ticks into the future, and the 2^N×2^N centers are stored and re-used without recalculating them, whenever the same large hashtiles show up again. This works because information cannot travel faster than the speed of light in Conway’s Game of Life and other rules of range 1: it is impossible for anything outside of the large tile to affect the center area in that amount of time.

  2. The message I take away from the game is that great (apparent) complexity can arise from simple rules. In this case just 4 totally dumb, simple rules create the appearance of considerable complexity.

    So, what would a million or so neurones create ? Emotions ? Creativity ?

  3. graham2:

    The message I take away from the game is that great (apparent) complexity can arise from simple rules. In this case just 4 totally dumb, simple rules create the appearance of considerable complexity.

    Yes. Who would have ever predicted, from those simple rules, that you could implement Life within Life as in the video above?

    So, what would a million or so neurones create ? Emotions ? Creativity ?

    It might take more than a million, but I agree with your point.

    Yet I’ve encountered plenty of dualists who argue that a system composed of neurons can’t implement things like creativity, emotions, or consciousness because the neurons themselves aren’t creative, emotional, or conscious. It’s basically a version of the fallacy of composition.

    I’ve seen Neil make a similar mistake by arguing that systems of logic gates can’t be used to implement AI, because the gates don’t possess the characteristics that would be required of the system as a whole.

  5. If you drop a bucket of paint on the ground, it will also make something that looks complex.

    I am pretty sure if you drop 1 million buckets of paint, it won’t become conscious.

  7. keiths: I’ve seen Neil make a similar mistake by arguing that systems of logic gates can’t be used to implement AI, because the gates don’t possess the characteristics that would be required of the system as a whole.

    Perhaps he was reasoning by analogy. Would evolution occur without mutations?

    Learning algorithms require a source of “randomness”.

    But the argument I’ve heard is based on performance. Current chip architectures can’t be scaled up to do real AI. Systems like Deep Blue require massive amounts of hardware just to compete with humans in fully deterministic games.

    Some philosopher should address the apparent paradox that brains are so much more energy efficient than silicon logic chips. And 100 Hz neurons so much faster than CPUs.

  9. I wouldn’t call what they’re simulating here in the video below “life”, because they don’t seem to be evolving in any way that is noticeable from the visualizations (it is not obvious to me that there is any selection or information transmission going on), but it is still fascinating to see how self-replication emerges from simple rules:

  10. Rumraket: …it is not obvious to me that there is any selection or information transmission going on…

    Indeed. The whole thing depends on initial conditions. There’s no possibility of variation being selected by the environment. See it once…

    ETA should have watched your video first. They vary the environment by increasing particle density.

  11. phoodoo:
    If you drop a bucket of paint on the ground, it will also make something that looks complex.

    I am pretty sure if you drop 1 million buckets of paint, it won’t become conscious.

    Your model lacks heritable traits, reproduction, and a competition for resources.

  13. T_aquaticus: Your model lacks heritable traits, reproduction, and a competition for resources.

    So if we drop enough paint, and pick the best ones each time, maybe eventually it will become conscious.

    Either that, or complexity doesn’t equate to sophistication.

  14. T_aquaticus: Your model lacks heritable traits, reproduction, and a competition for resources.

    Do you think any example that includes all of those must have some sort of functioning….I don’t know….elan vital or something like that? A “desire” to propogate or “get it on” or suchlike? A ‘Prime Directive,” as it were?

    Could a line in computer code be sufficient?

  16. phoodoo: So if we drop enough paint, and pick the best ones each time, maybe eventually it will become conscious.

    The paint isn’t competing for resources, isn’t reproducing, and doesn’t have heritable traits.

  17. walto: Do you think any example that includes all of those must have some sort of functioning….I don’t know….elan vital or something like that? A “desire” to propogate or “get it on” or suchlike? A ‘Prime Directive,” as it were?

    Could a line in computer code be sufficient?

    Those that have heritable genetic traits that cause them to reproduce will quickly outnumber those that don’t.

  19. T_aquaticus: The paint isn’t competing for resources, isn’t reproducing, and doesn’t have heritable traits.

    But this silly program is?

    And what do we get..pixels. Lots of pixels. Ooooo!

    Will it gain consciousness one day?

  20. Even Conway himself says he doesn’t like his little game much.

    But he does say something rather preposterous and dim.

    He says, “Who would have thought you could get such a decent simulation of real living things in such a simple way.”

    I am not sure though if that is his original idea or if he just repeated it, after hearing it from Snooki on Jersey Shore.

    Patterns blink on and off depending on the number of pixels next to it. Exactly like life! Virtually identical.

  22. ‘Tetris’ Recreated In Conway’s ‘Game of Life’

    In 1970 mathematician John Conway created rules for the “Game of Life,” a now famous “zero-player game” where a grid of cells evolves (following Conway’s rules) from an initial state proposed by the player. In 2013 someone challenged readers of StackExchange’s “Programming Puzzles & Code Golf” section to devise an initial state “that will allow for the playing of a game of Tetris.”

    An anonymous Slashdot reader reports that “This challenge sat around, gathering upvotes but no answer, for four years. Then, it was answered.” Citing the work of seven contributors, a massive six-part response says their solution took one and a half years to create, and “began as a quest but ended as an odyssey.” The team created their own assembly language, known as QFTASM (Quest for Tetris Assembly) for use within Conway’s mathematical universe, and then also designed their own processor architecture, and eventually even a higher-level language that they named COGOL. Their StackExchange response includes a link to all of their code on GitHub, as well as to a page where you can run the code online.

    One StackExchange reader hailed the achievement as “the single greatest thing I’ve ever scrolled through while understanding very little.”

  23. At the other extreme:

    Tetris in 140 bytes

    mikejuk writes
    “Is it possible to write a JavaScript program in no more than a tweet’s length? A website called 140byt.es says it is and has an implementation of Tetris to prove it. Ok, it only has two types of block — hence its title “Binary Tetris” — and there’s no rotate, but it works. The blocks fall down the screen and you steer them into place. You can try it out by playing the demo. Of course the real fun is in figuring out how it works and there is lots of help on the site — so if you’re bored how about the 140 character challenge?”

Leave a Reply