0505: "A Bunch of Rocks"

[GooglyBoogly]

Where does "Oh shi-" come from?

WARNING! LINK NOT SUITABLE FOR MINORS
http://www.encyclopediadramatica.com/Divide_by_0

Should give you some reference


And now to more pressing matters. This comic reminded me of a quote from Sid Meirs: Alpha Centauri - The turn based- strategy game

We are all aware that the senses can be deceived, the eyes fooled. But how can we be sure our senses are not being deceived at any particular time, or even all the time? Might I just be a brain in a tank somewhere, tricked all my life into believing in the events of this world by some insane computer? And does my life gain or lose meaning based on my reaction to such solipsism?

Project PYRRHO, Specimen 46, Vat 7
Activity Recorded M.Y. 2302.22467
TERMINATION OF SPECIMEN ADVISED

It sounds much cooler in the .wav attached here

[atimholt]

I know I'm overthinking this (as if that were possible, with this premise), but I wouldn't bother laying rows of rocks in a pattern. Before one could finish laying a complete row, I bet I could study my strange new universe to a tee, smelt surrounding raw materials, build robots and complex computers, then synthesize life and get some real human company. Before one could lay a single row, I would be an immortal god of a real people. My empire would span light years (maybe, I'd have to do the math) and the land would be lush. If there were too little water, I'd transmute my own hydrogen and oxygen, and would be able to afford to wait a billion years for it to lose any radioactivity it might have if that were the only way. A little tiny bit like Riverworld, in that whole making-amazing-stuff-starting-with-very-little sort of way, but I'd have more time to refine, and I'd probably have more metal.
For curiosity's sake, one thing I'd do is rocket up to as high an altitude as I could. What would I see from a million miles up, homogenous brown expanse? A scrawled message?
And maybe I'd dedicate a certain slice really far away to an ever expanding computer.
But don't get me wrong, the rock + lonelyman computer is still an awesome concept.

EDIT: More time than those of Riverworld, what am I talking about? Less distraction and more metal would be my main advantages. By the way, I don't think I really liked Riverworld, though I read the four books of the original 'trilogy'. Okay, I didn't like all of it. I liked some of it.
You guys know what I'm talking about, right?

[Zachariah]

And does my life gain or lose meaning based on my reaction to such solipsism?

But being a brain in a tank, or a programmed simulation, is the complete reverse of solipsism. It's the realisation that you are completely detached from the consequences of your actions - the sort of freedom only true sociopaths get to experience. It's simlar, in a way, to the fatalism some religions engender - why take any safety precautions when it is the will of god if you get hit by a bus?

[Fratboy]

Now, me not being a programmer, i would have just made a nice little society out of all those rocks...
i guess i'm just that simple.

[Hiato]

Some here raised this point earlier, in a kind of vague sense (that is, I don't think it was intended), but [there always is one]:
What qualifies existence? Trivially we cannot prove that anything, save ourselves, exists. The entire universe and everything around us may, in fact, just be a figment of our imaginations. There is further instinctive logic in this: If nature settles on the most efficient solution, then existence sure isn't it. Imagine this, we have an individual who imagines himself, now ostensibly, if this individual truly exists, then he has the power to imagine himself wholly and thus the imagination has the power of imagination. Now, the imagination merely has to imagine another one of itself. What we find is that once we exist, the most efficient solution would be to imagine ourselves again, and then cease existing totally. In that instant, the created self can create another one of itself (ad infinitum) who is not dependent on the previous one as it can too, create another one of itself. Thus, following this logic, existence is pointless, non-existent (for the lack of a better word) and a trivial property as nothing actually exists in this theory. [Of course, somehow, there would need to come into existence the very first being capable of spawning this chain - however, the instant the chain arose, he would no longer be needed]

On the other hand, if only "real stuff" (as defined by some property) exists, then we come across the problem of self-evidence. If stuff only exists because we think it does, then who's to say that that property, being "real stuff", exists in the first place. Why can't I just simulate something and then, by virtue of me being "real stuff", it becomes "real stuff" too as it can be directly traced to the actions and effects that "real stuff" has.

Both of these paths, however, tend to lead us to the same conclusion. It is both impossible to prove whether anything exists and impossible to deny it - that's not a very nice conclusion though

(I hope this doesn't violate first post rules. EDIT: Typos and more)

[Aikanaro]

Okay, sorry, but put me in with the people who find this comic depressing and sad as all hell. Not because of the ramifications for free will, but because I just feel sorry for Desert-Guy. Some folks have been comparing this to Small Gods....anyone remember what happened to Vorbis? The human mind would snap like a brittle twig after a couple YEARS of isolation....after a century, you'd be catatonic. I'd personally view his existence as a form of Hell. Not the WORST form, to be sure....but even so, I'm surprised the guy never, in all those eons, just tried to bash his own brains out on one of those rocks.

As for the computing, that also requires the assumption that Desert-Guy has a near-perfect memory with unlimited capacity, and an attention span to match. Otherwise, by the time you started coding a line for the rockputer, you'd forget where you were on it, what the line was supposed to do, etc.. I mean, let's say the code takes a space of a mile....you can't SEE that far, while being able to clearly make out how far along you were. You'd forget which section of it you were working on. Thus, this removes even the "universe simulator" as a good way of keeping yourself occupied.

[John Sidles]

Three good ways to learn more about the mathematics behind this (outstanding!) comic are:

• visit http://rendell-attic.org/gol/tm.htm
• Google for the exact string "JHWH Conway"
• Read Volume 2 of Winning Ways (for your Mathematical Plays)

When it comes to simulating a large quantum universe (as contrasted with a large classical universe), this is thought (by some folks) to be PTIME computationally feasible only for noisy (finite-temperature) universes ... which our universe happens to be!

GPL'd simulation software for noisy quantum systems (slightly faster than rocks) is here: http://faculty.washington.edu/sidles/QSEPACK/

---- update edit ---
Cool. Googling for the exact string "JHWH Conway" now leads to this xkcd forum.

The point being, that evidently it's the Google Corporation that's running our simulation.

Well, that sure explains a lot!

... and I, for one, welcome our new Google rock-indexing overlords!

[Mazzula]

I suggest looking further into Mealy Finite State Machines. By definition, they need to know their current state right "now".

I think you are being mislead by the name "current state" as used in this context. The current state of a state machine doesn't have anything to do with what is happening right now. It just refers to some internal value in the software that is running the state machine, a value that is part of the overall machine state. It is completely sensible to ask what the state of the machine was yesterday at noon. We would still be talking about the "current state" even though we were talking about a time in the past.

The Mealy FSM doesn't really bring anything extra. It is just a matter of how the machine is specified. In the Moore model outputs are a function of the state alone, whereas in the Mealy model the outputs are a function of both the machine state and the state of the inputs. So the Mealy model can be considered as a Moore model where the states are taken as a cross product of the Mealy state and the input states.

Mathematically, the Finite State Machine is fully specified without regard to the actual inputs that are presented to it in some situation. These inputs come into the machine and are represented in some internal form, perhaps as bits in some hardware register. But as far as the software is concerned that is running the machine, it is all just internal machine state, and the operation of the software is just a matter of a sequence of transitions of this internal machine state.

Let's get back to the comic. Suppose there is some row of rocks which includes a transition of a Mealy state machine, maybe the noon transition. So long as that row of rocks existed, unchanged, the noon transition would have happened.

Further, consider that the 1PM row of rocks is entirely independent of the noon row of rocks. So, in the model of the rocks, any evidence we have at 1PM that the state machine made the noon transition, is contained in the 1PM row of rocks. That is, the 1PM universe doesn't have access to the noon row of rocks. If Randall went back and changed the noon row of rocks, we (that is, our 1PM selves) wouldn't know it.

[jbakid]

I think that he meant rule 30 (http://en.wikipedia.org/wiki/Rule_30). Rule 34 is just a diagonal line: (Here's a program in C to sort stuff out: to do rule n, and write to a file filename.ppm that has height h pixels, width w pixels, just use the command ./cautomata w h filename.ppm n 3. The 3 is to say how many bits are read in. (Of course, we all know what rule 34 of the Internet is)

Code: Select all

// cautomata.c
#include<stdio.h>
#include<stdlib.h>
int main(int argc, char** argv){
  int w = atoi(argv[1]);
  int h = atoi(argv[2]);
  FILE* f = fopen(argv[3], "w");
  char* a = calloc(w, 1);
  char* b = calloc(w, 1);
  char* x;
  int i,j, rule = atoi(argv[4]);
  int r, c, k = atoi(argv[5]);
  a[w/2] = 1;
  fprintf(f, "P3\n%d\n%d\n255\n", w, h);
  for(i=0;i<h;++i){
    for(j=0;j<w;++j){
      if(a[j])
        fprintf(f, "%d %d %d\n", 255, (255 * i) / h, (255 * i) / h);
      else
        fprintf(f, "%d %d %d\n", 0, 0, (70 * i) / h);
      c = 0;
      for(r = 0; r<k; ++r)
        if(r+j-k/2 >= 0 && r+j-k/2 < w)
          if(a[r+j-k/2])
            c += 1 << (k-r-1);
      if(rule & 1 << c)
        b[j] = 1;
      else
        b[j] = 0;
    }
    x = a;
    a = b;
    b = x;
  }
  fclose(f);
  return 0;
}


Oh, and you'll need to have the command "convert" installed on your computer if you do not have a .ppm viewer.

[jbakid]

I think that he meant rule 30 (http://en.wikipedia.org/wiki/Rule_30). Rule 34 is just a diagonal line: (Here's a program in C to sort stuff out: to do rule n, and write to a file filename.ppm that has height h pixels, width w pixels, just use the command ./cautomata w h filename.ppm n 3. The 3 is to say how many bits are read in. (Of course, we all know what rule 34 of the Internet is)

Code: Select all

// cautomata.c
#include<stdio.h>
#include<stdlib.h>
int main(int argc, char** argv){
  int w = atoi(argv[1]);
  int h = atoi(argv[2]);
  FILE* f = fopen(argv[3], "w");
  char* a = calloc(w, 1);
  char* b = calloc(w, 1);
  char* x;
  int i,j, rule = atoi(argv[4]);
  int r, c, k = atoi(argv[5]);
  a[w/2] = 1;
  fprintf(f, "P3\n%d\n%d\n255\n", w, h);
  for(i=0;i<h;++i){
    for(j=0;j<w;++j){
      if(a[j])
        fprintf(f, "%d %d %d\n", 255, (255 * i) / h, (255 * i) / h);
      else
        fprintf(f, "%d %d %d\n", 0, 0, (70 * i) / h);
      c = 0;
      for(r = 0; r<k; ++r)
        if(r+j-k/2 >= 0 && r+j-k/2 < w)
          if(a[r+j-k/2])
            c += 1 << (k-r-1);
      if(rule & 1 << c)
        b[j] = 1;
      else
        b[j] = 0;
    }
    x = a;
    a = b;
    b = x;
  }
  fclose(f);
  return 0;
}


Oh, and you'll need to have the command "convert" installed on your computer if you do not have a .ppm viewer.

Add'l info:
http://mathworld.wolfram.com/CellularAutomaton.html
http://mathworld.wolfram.com/Rule30.html
http://mathworld.wolfram.com/Rule110.html
http://mathworld.wolfram.com/Rule90.html
http://mathworld.wolfram.com/Rule126.html
and finally http://mathworld.wolfram.com/topics/Cel ... omata.html

[ThemePark]

I think that he meant rule 30 (http://en.wikipedia.org/wiki/Rule_30).

Although you and a few others have pointed this out, I highly doubt he made a mistake. He is not referring to a specific Wolfram pattern, just making a pun on the number 34. Saying "I call Rule 34 on Wolfram's Rule 30." wouldn't make any sense, unless Rule 30 or any other Rule for that matter looked very much like a sex position as is.

[GCM]

I know I'm overthinking this (as if that were possible, with this premise), but I wouldn't bother laying rows of rocks in a pattern. Before one could finish laying a complete row, I bet I could study my strange new universe to a tee, smelt surrounding raw materials, build robots and complex computers, then synthesize life and get some real human company. Before one could lay a single row, I would be an immortal god of a real people. My empire would span light years (maybe, I'd have to do the math) and the land would be lush. If there were too little water, I'd transmute my own hydrogen and oxygen, and would be able to afford to wait a billion years for it to lose any radioactivity it might have if that were the only way. A little tiny bit like Riverworld, in that whole making-amazing-stuff-starting-with-very-little sort of way, but I'd have more time to refine, and I'd probably have more metal.
For curiosity's sake, one thing I'd do is rocket up to as high an altitude as I could. What would I see from a million miles up, homogenous brown expanse? A scrawled message?
And maybe I'd dedicate a certain slice really far away to an ever expanding computer.
But don't get me wrong, the rock + lonelyman computer is still an awesome concept.

EDIT: More time than those of Riverworld, what am I talking about? Less distraction and more metal would be my main advantages. By the way, I don't think I really liked Riverworld, though I read the four books of the original 'trilogy'. Okay, I didn't like all of it. I liked some of it.
You guys know what I'm talking about, right?

O yea... I can totally smelt raw materials with my bare fists too.

[Electric Prophet]

Now, me not being a programmer, i would have just made a nice little society out of all those rocks...
i guess i'm just that simple.

Same here. Except that last panel was GOLD.

[*Wildfire*]

No, it looks much more like 110 than 30 to me...

I know this is trivial, but its math; there is a right and wrong answer, and 30 is the former, 110 the latter. If you look at all the rules, several emerge into similar, yet differently orientated patterns. 60, 102, 110, and 126 all have the triangles, but are too uniform to be the one in the comic (there are properties other than simply containing triangles). Notice how the edge is very uniform, but then as you move right the triangles begin to get more random in size. That is most certainly 30. Conway's favorite, apparently, because of its repeating, yet seemingly random nature.

P.S. 80 goes the other way.

http://mathworld.wolfram.com/Elementary ... maton.html

[DeadCatX2]

I suggest looking further into Mealy Finite State Machines. By definition, they need to know their current state right "now".

I think you are being mislead by the name "current state" as used in this context. The current state of a state machine doesn't have anything to do with what is happening right now.

I think we're just using different definitions of "current state". The current state of a state machine represents what is happening right now. The person who designed the FSM intended for some variable to represent what is going on "right now", so by using this variable computers certainly have a sense of "right now".

Let's get back to the comic. Suppose there is some row of rocks which includes a transition of a Mealy state machine, maybe the noon transition. So long as that row of rocks existed, unchanged, the noon transition would have happened.

Further, consider that the 1PM row of rocks is entirely independent of the noon row of rocks. So, in the model of the rocks, any evidence we have at 1PM that the state machine made the noon transition, is contained in the 1PM row of rocks. That is, the 1PM universe doesn't have access to the noon row of rocks.

I do not see how a lack of history has anything to do with computers knowing the current state. For that matter, we could design a computer to record all state transitions and we could then retrieve the "noon transition". This leads me to believe we are talking about different things. How do you get from

That is, a computer program doesn't know what its current state is, it only knows the state as a function of some index, but it has no way of knowing what the value of the index actually is.

to computers only knowing previous states as a function of the current state?

I'm not trying to prove you wrong, I'm trying to understand your point. To me, the bits that represent the current state of an FSM are how it knows what the current state is. Without that value, you could not make an FSM.

The Mealy FSM doesn't really bring anything extra. It is just a matter of how the machine is specified. In the Moore model outputs are a function of the state alone, whereas in the Mealy model the outputs are a function of both the machine state and the state of the inputs. So the Mealy model can be considered as a Moore model where the states are taken as a cross product of the Mealy state and the input states.

Right, I should have generalized to FSMs, Mealy or Moore. For some reason, I had it in my mind that Moore machines made transitions based on inputs (which is silly in hindsight, that's just plain old combinational logic), and Mealy machines made transitions based on inputs and current state. Regardless of my error, both models require some knowledge of what the current state is, so I think your use of "current state" comes with a connotation that seems inappropriate to me. I also don't think your use of "cross product" is typical for this field.

[Endless]

What does he do if he finds a watch?

[EmpJohnIV]

me: if he has finite memory, I doubt he can even remember when or where he began.
Carter: this is a very good way to explain to people how vast arrays of determenistic processes can do incredible things
Carter: well if he translates each line into a state as he goes he could write out, in language, or draw, or whatever, whats going on he could make notes translate bit by bit find out what happened I guess then with enough time and patience he could find out a lot of what took place and time and patience is what this is about
me: not really the process of decoding would take more rules then I think he can remember
Carter: it depends on the coding more than anything
me: maybe
Carter: and how localized things are he would be the compiler
me: so the universe is a huge triangle, a trillion trillion trillion trillion trillion rows of stones.
Carter: and it might take many stages of compiling to get something that was readable
yes
me: He decides to pause and look in to a current state. He goes to the side and makes a copy of the rows needed to represent the current state he goes to a place in the desert where he wrote out the direction for decompiling, loads those directions in to active memory (memorizes them)...
and uses the code to solve find out about the current state of the universe
Carter: this is a move from binary on up
me: across the desert are many rock gardens with directions telling him how to go to the next step done as he derived the begining of his process for loading in to active memory at the time needed
Carter: or a map carved in a rock he carries?
me: no too much information to carry.
Carter: ah
me: most of the rock gardens are direction to which rock garden to go to and what to do when getting there
like memento. He doesn't know where he is going. On this time scale human memory is so limited
Carter: long, long walks
me: the directions tell him to set up the gardens that will one day tell him to use them after a long time of following directions, that he does not remember writing but knows that he did indeed write, he follows a derrivation that he finds on one extreme of his explored realm
and he decodes this huge huge huge line of stones, older then can be conceived, to derive them into something that is human readable. after eons of decoding he walks to the beginning of the text.
Our Father, who art in heaven,
Hallowed be thy Name.
Thy kingdom come.
Thy will be done,
On earth as it is in heaven.
Give us this day our daily bread.
And forgive us our trespasses,
As we forgive those who trespass against us.
And lead us not into temptation,
But deliver us from evil.
For thine is the kingdom,
and the power, and the glory,
for ever and ever.
Amen.
Carter: self portrait?
me: maybe
or if he is lucky the text may read
'you're welcome'. But what ever it says, it will end in a map to the triangle of stones 10^70 light years away, go on finish the story.
Carter: ha ha
me: Sisyphus has nothing on this guy.
Carter: i suppose the translating of language of whatever life occurs adds a humongous amount of work
particles moving about is not so bad, but it takes a lot of them....
me: a googol years ago he forgot what he was doing, now he just moves stones out of habit, and he chants to him self the content of that comic (more or less) although the words long, long, long ago lost meaning.

[Endless]

me: if he has finite memory, I doubt he can even remember when or where he began.
Carter: this is a very good way to explain to people how vast arrays of determenistic processes can do incredible things
Carter: well if he translates each line into a state as he goes he could write out, in language, or draw, or whatever, whats going on he could make notes translate bit by bit find out what happened I guess then with enough time and patience he could find out a lot of what took place and time and patience is what this is about
me: not really the process of decoding would take more rules then I think he can remember
etc.

The comic made my day. Adding this conversation made my week.
And Rock-man made my eternity.

[John Sidles]

As a followup to the above post, I went to Paul Rendell's Life Page and downloaded his Turing Machine (in effect, a very special XKCD rock-pattern that emulates a Turing machine).

Then I downloaded Andrew Trevarrow's and Tomas Rokicki's wonderful cross-platform Life simulator Golly



Hurrah! ... Golly implements Rendell's Life-Turing machine perfectly! And it is mighty fun to watch in action (for a quantum system engineer like me).

Next, inspired by Gregory Chaitin's work, the obvious next step was to launch a universal Life meta-machine ... meaning, a Life constructor machine that systematically constructs a sequence of universal Turing machines, with each successive machine initialized with the next Turing program ... as synthesized in (what else) alphabetical order (actually, this capability is not yet available ... Paul Rendell's amazing page describes ongoing work in this direction).

Gosh, I wonder what fraction of these Turing programs will ever stop? Thinking about this question led a (then-teenage) Gregory Chaitin to amazing mathematics.

What all this means of course (in practical terms) is that the whole XKCD forum (including me, I guess) is now already being simulated ... on an ordinary laptop ... provided that we keep upgrading the memory, and no one shuts off the power.

So, heck ... there's no need to wait for Wednesday's XKCD comic ... or read the forum posts about it ... or even wonder about your inner thoughts about this post ... `cuz we've already started to simulate all of them! ... with no need to wait for an infinitely wise programmer!

Gosh, which Turing stack will these results (namely, the simulations of us) be in?

[Mazzula]

The person who designed the FSM intended for some variable to represent what is going on "right now", so by using this variable computers certainly have a sense of "right now".

If you want to define the term "right now" to be the value of that variable, then I agree with you. However there is nothing that links the value of that variable to the usual subjective idea that a particular time is "right now".

To me, the bits that represent the current state of an FSM are how it knows what the current state is. Without that value, you could not make an FSM.

For example, you might simulate the operation of that state machine by writing values down on successive sheets of a memo pad. (Or successive rows of rocks.) For every transition, you turn the page or move to the next row of rocks and represent the next state there. If you do that, you will see that the operation of the state machine, in some particular circumstance, is actually just a sequence--a mapping from page number in the notepad (or rock row number) to state.

So where does the notion of "now" enter into it?

The way it enters is always by moving to something outside the mathematical model--some human perception. There is nothing in the particular page of the memo-pad simulation, nor in the contents of the row of rocks in the desert, that distinguishes that page or that row as representing "now" within the simulation.

I also don't think your use of "cross product" is typical for this field.

You are correct, I should have written "Cartesian product". Sorry about that.

[TheWhattabatt]

I bet he'll be ticked when he realizes the stones were actually crabs waiting for the right time to scuttle away.

Similarly: Jack Sparrow is a massive underachiever.

[atimholt]

O yea... I can totally smelt raw materials with my bare fists too.

Remember, you've got an infinite amount of time to be clever. At the worst, I'd build a mountain of rocks and work out some kind of friction smelter working on sheer dropping force. (or mountainous pressure!)

[Kow]

Panel #15: String Theory anyone?

[cantimatt]

Please make this available as a poster print I want to hang this in my office!

[phlip]

I know this is trivial, but its math; there is a right and wrong answer, and 30 is the former, 110 the latter.

Just to check, we're talking about the same rule 30 and rule 110, right? And the same comic? 'Cause the comic is very clearly not 30. For one thing, the hypotenuse of the recurring triangles is at 45 degrees, right angle in the top left, whereas rule 30 has a horizontal hypotenuse, right angle at the bottom.

More to the point... the pattern is rule 110. Here's one bit of it, with the 8 different conditions highlighted... note how they exactly match rule 110.

[Clintonio]

I'm no computer science student, but I've delved into this area before, and have long considered a possibility like this.

But then I realised that it doesn't affect my day to day life and probably got hammered that evening after finishing a maths assignment.


Probably the best XKCD yet. So much conveyed in such a simplistic strip.

[Diadem]

As for the computing, that also requires the assumption that Desert-Guy has a near-perfect memory with unlimited capacity, and an attention span to match. Otherwise, by the time you started coding a line for the rockputer, you'd forget where you were on it, what the line was supposed to do, etc.. I mean, let's say the code takes a space of a mile....you can't SEE that far, while being able to clearly make out how far along you were. You'd forget which section of it you were working on. Thus, this removes even the "universe simulator" as a good way of keeping yourself occupied.

He doesn't have to have a very good memory. You're assuming the new state of the universe after one clockcycle is dependent on the full state of the previous cycle. That's not necessary, and in fact extremely improbable. Probably the next state of every single cell in the grid that makes up the universe only depends on the previous states of itself and its neighbouring cells. That's actually how cellular automata work as well.

For our universe, this is in fact needed. Communication is capped by the speed of light after all. And if our universe is discrete, then probably the speed of light translates to moving 1 cell each timestep

[theta4]

Oh, I would so hate to be that guy. It would get so boring after such a long time.

Also, I don't know if the question has been answered, but there is no visual representation of the simulation. If there are people in that simulated universe observing the universe, they are able to do so because the simulated light waves bounce off of the simulated particles, and enter simulated eyes to send simulated messages to a simulated brain. Of course, the simulated brain has absolutely no clue that it's being simulated. So it appears to the brain like it's a reality. The poor guy will never get to see his simulation unless he finds a way to make a representation of the data. All that matters for any observation within the universe itself is that the physics work properly. It gives you a whole new perspective, doesn't it?

I hope someone reads this.

It makes you wonder if we are simulated...how would we know? The entire universe is like a big quantum clock (according to our understanding), and with enough time, it could be simulated!

Neurons are simulated. Simulated neural networks can learn.

What would happen if a brain were simulated? If it were simulated down to the neuron, what would it be like? The simulated brain would of course be conscious.

What if the body was simulated? If it were simulated down to the cell, what would it be like? The simulated person would be conscious, and would be able to act and learn in its environment. If you applied gravity, it would feel gravity. It would be incredible. What if you were in the shoes of that person? You would only see what you were able to see. That is, whatever enters your simulated eyes. The only way that the person would know they are being simulated is if you told them that you made them.

The body would work like a normal body, regardless of who is observing it. In fact, nobody needs to observe it. The only thing that matters in that the simulation keeps running.

--alex

I hope that helps with your understanding of how I see this.

[RanCorp]

...
For our universe, this is in fact needed. Communication is capped by the speed of light after all. And if our universe is discrete, then probably the speed of light translates to moving 1 cell each timestep

Which is consistent with the values for the Planck time and distance (as given in the table at http://en.wikipedia.org/wiki/Planck_units):

• Planck Time: 5.391x10⁴⁴ m
• Planck Length: 1.616x10^-35 s

L_planck / T_planck ~= 300,000,000 m/s

(that's a tilde before the equals—meant to denote approximate equality—in case your browser font makes it look like a minus sign, as mine does.)


RanCorp

[phlip]

Which, given that the Planck units are designed such that c, among many other constants, is 1, shouldn't be surprising.

[Aquila]

First of all, I don't know that it would get boring. I can't imagine I'm the only one who gets wrapped up in a project and finds eating and sleeping to be a bother. I think a situation like that would be interesting in many ways. Nothing to distract me from my own thoughts. That said, I think it's impossible for any of us to imagine what eternity is. You would have to come up with something to occupy your time and energy or you'd go nuts.

Secondly, as far as visualization goes, it would be nice but unnecessary to have something to view the simulation. Again, I wouldn't consider myself unique in the ability to "read" my code and "see" what it does. Granted, I've never done anything near this complex. But with infinite time, and being the original designer of the system, I would imagine you'd have a lot more familiarity with it and be able to do much the same thing.

[fishyfish777]

I have converted from frisbeetarian (George Carlin) to Randallism (Randall Munroe)

[TinBromide]

Rule 110 is the only Turing complete wolfram's rule. (well, it and its inverses and reflections) Rock guy is using rule 110. Also, the "program" appeared to generate large triangles, you think that might be due to the programmatic nature of the universe?

IIRC rule 30 has class 4 behaviors (semi-random output, like 110), but 110 was proven to be capable of being a universal turing machine.

You can do and gates, or gates, not gates, and from the combination, all the "circuits" one would need to build a computer. Granted that each bit would be made up of hundreds or thousands of stones and each particle might be miles apart from each other, but eventually, they might cause systems to interact. That being said, the thought of running a rule 110 turing machine on my computer seems so incredibly recursive in light of this comic...

As for the person who mentioned having to remember what happened miles down the line, each rock (bit) in the line is determined by the state of the 3 in the row previous (arranged in a T tetramino where the bar up top is the 3 input bits and the bottom is the output).

The thing is that to handle a single iteration might take billions of billions of billions of years to even evaluate every bit in the next line. Its crazy stuff. However, what rock guy is doing right now is examining 3 rocks deciding if the output bit gets a rock or not, not exactly engaging stuff. I wonder if he's capable of reading the output and theoretically knows everything there is to know in the universe at every point in time...

[EmpJohnIV]

Unless we can assume that desert man has an arbitrary amount more memory then us, he wouldn't be able to read much of the universe at all. The number of stones needed to capture the event of a bacteria dividing would cover many billion square light years. This is an awesome code, but not a good one to read.

I wonder if there can be a pattern like Wolfram 110 that can handle fuzzy logic? Where the size of the rocks, or their orientation can matter. I have been thinking about the opinion that our universe is not distinct (which I have no knowledge ether way), a continuum, something that does not 'tick' forward. I have a gut feeling that adding fuzzy logic could help with this problem and even some of the quantum weirdness.

The initial state that Desert guy set up must be huge, many times more stones then fundamental particles. Stretching on for at least 10^50 light years, most likely this is a shy estimate, by a lot. How could he set it up at all being human and what have you? Well as far as we can tell the initial state of the Universe was relatively uniform, repetitive. He figures out all the forces, and types of particles, and the codes for each, the is an absurdly large amount of information, but a less then there will be shortly. The information is insanely repetitive, but now it is a patter that he just follows. Once he has line one for a perfectly consistent universe set up, he need only change a tiny tiny portion of the variables and chaos theory will create a universe that does its own thing. Unlimited complexity can spring forth from just a little bit of chaos. Of course he can't know what is happening in his Universe (assuming the desert man has limited enough mental capacities to justify dropping stones to begin with) but thats fine. The program need not be viewed, because we are in the program, and what appears as consciousness are strange loops. The program views itself.

And if you are a materialist (our universe is made of stuff, and not virtual stuff) then what is all of this (wide arm gesture) other then the derivations of a computer made out of {insert physics stuff here}. I mean, a table is the same thing as a representation of a table, as rendered by whatever our universe is at base. Representation is a bad word to use here, as it presupposes the old Cartesian dualism I think, you have mind and material world, the mind has a representation of (blank), but doesn't representation sound a little like copy, like it suggests the mind is one thing and material universe external, and the mind holds a copy (good copy of crummy copy) of the matter bing represented. Lets say that I have an idea of a tree under this view. My mind holds or maybe views (depending on the metaphor family you want to use) a representation of the tree, a model, a copy like thing. But isn't that view crazy? My mind doesn't have a representation of a tree, it is a lets not say representation of the tree. Look at how ridiculous the Cartesian epistemology is: How do I see a tree? You mind sees an mental image of the tree. How do I see that? Oh shi-.

So going back to the start of this I don't mean that the universe is representing the table in anyway more then to say really that the universe is being a table (at least partially of course) so some of the universe is being a table. When a large group of computer circuits are rendering a web page, aren't they being a web page, or some would say the stuff of a web page?

But is this stuff of the universe really what we think it is as opposed to stones? John Locke described stuff (substance) and 'something, I know not what' because it is not an observable property, it is that in which properties dwell. But we can never get to substance only to more obscure properties, the "thing in itself" is beyond our perception. Maybe we can get to the stuff of a web page, a computer, silicon, quarks, strings, e8. What is the stuff of strings, or energy, or e8? So, is the universe really made of strings, particles, stones, dreams, ideas, Turing machines or any other foundational, basic thing? Well, what is that foundational thing made out of? Once the flat earth was criticized with the phrase "Turtles, turtles all the way down." Do we face the problem of "Turing, Turing all the way down"?

[MrSparkle]

So, is this a simulation of the universe minute by minute or is it just meant to run a simulation if the stones were read?

And can he ever, really, be done?

To simulate the universe in real time you'd need a quantum computer. Then again, desert guy seems a lot smarter than I.

You would need a quantum computer the size of the universe. You see, the universe actually is a quantum computer, the question is: What is it calculating?

[scarletmanuka]

Although Greg Egan's Diaspora and Permutation City have already been brought up, allow me to draw your attention to another of his novels, Schild's Ladder, which also has some good stuff on this subject.

The Swiss patent office reference implies that desert guy is Einstein.

I didn't read it that way. I read it as him choosing Einstein as a comparison that would be meaningful to us.

It's great to see that this idea is spreading. Imagine having an infinite, but empty HDD. What would you do with it? Start filling it of course, would be a waste otherwise. Fill it with what?

Well, porn, of course. Duh. And probably music and games too.

Now that's real programming.

Perhaps we can appeal to the guy in the desert to go back a few decillion rows and alter the universe so that our own desired programs naturally appear on our hard drives?

... Except that his definition of "naturally appear" probably correlates to "programmer designs code, implements it, tests, fixes bugs, finds out what the real requirements were, etc". Oh well.

I wonder if there can be a pattern like Wolfram 110 that can handle fuzzy logic? Where the size of the rocks, or their orientation can matter. I have been thinking about the opinion that our universe is not distinct (which I have no knowledge ether way), a continuum, something that does not 'tick' forward. I have a gut feeling that adding fuzzy logic could help with this problem and even some of the quantum weirdness.

I don't know if the fuzzy-logic thing would help the guy in the desert much. Sure, he'd have fewer rocks in each row, but imagine the time he'd have to spend shaping each rock to fit the requirements (or worse, trying to find a rock of exactly the right size and shape) for each spot.
As for whether it's possible as a concept, if you make it discrete (say, allow only 8 different categories of rock size) then it's essentially just a packing scheme and can be implemented as a cellular automaton, though you would then have to allow cells to depend on more than their immediate neighbours (in the unpacked view). If it's continuous, so that the rules are along the lines of "If there is a rock of size x at position i-1, and a rock of size y at position i, and no rock at position i+1, then the result for position i is a rock of size f(x, y)", then I don't think you could do this.

The whole concept of a cellular automaton depends on a time 'tick'. If you're simulating a universe with it, I don't see that it necessarily follows that the simulated universe must also have a time 'tick', but I find it very hard to imagine how it could not.

[RanCorp]

...
You would need a quantum computer the size of the universe. You see, the universe actually is a quantum computer, the question is: What is it calculating?

The evolution of the quantum-mechanical wave function of this universe's physics with initial boundary conditions equal to those of this universe's big bang.

What else?

RanCorp

[Deaths_son]

the latest xkcd is... dammit. we know you understand how computers work already! this does not make for a joke! I KNOW you have the potential to make me laugh. please do so!

[neurosci_queen]

How would Rule 34 even APPLY to Wolfram's rule?? I mean, waht kind of porn could you make of that??!?!


I'm seriously asking here, guys.

[EmpJohnIV]

I don't know if the fuzzy-logic thing would help the guy in the desert much. Sure, he'd have fewer rocks in each row, but imagine the time he'd have to spend shaping each rock to fit the requirements (or worse, trying to find a rock of exactly the right size and shape) for each spot.
As for whether it's possible as a concept, if you make it discrete (say, allow only 8 different categories of rock size) then it's essentially just a packing scheme and can be implemented as a cellular automaton, though you would then have to allow cells to depend on more than their immediate neighbours (in the unpacked view). If it's continuous, so that the rules are along the lines of "If there is a rock of size x at position i-1, and a rock of size y at position i, and no rock at position i+1, then the result for position i is a rock of size f(x, y)", then I don't think you could do this.

The whole concept of a cellular automaton depends on a time 'tick'. If you're simulating a universe with it, I don't see that it necessarily follows that the simulated universe must also have a time 'tick', but I find it very hard to imagine how it could not.

Very interesting, I was looking for a way to get around the problem of discrete computation. I think maybe it can be avoided, with rocks of a spectrum of weights, where weight of the rocks from the higher row affects the rocks on the lower row. And a rock may be skipped if it is below a threshold. But I am really talking out of my butt at this point, I don't know much about fuzzy logic. I can't help thinking that a fuzzy logic system could be used, although perhaps it would make sense in a different fantasy setting.