The Naked Scientists

The Naked Scientists Forum

Author Topic: A computer simulation that ccan predict the future  (Read 5277 times)

Offline realmswalker

  • Sr. Member
  • ****
  • Posts: 205
    • View Profile
I have been thinking and philosophizing and contemplating this idea for sometime now.
given the exact same starting conditions, a situation will always end up with the exact same outcome, correct?
Well when the universe was formed, it existed, before the big bang, as a  Homogenous ultra dense substance, there was no variation in it. After the big bang that substance went on to create everything, and we right now exist because of that reaction.
So, logically, if you took the exact same quantity of that homogenous material, the true primordial soup, and set of a new reaction with it, a universe identical to ours would form. It would have the exact same starting conditions, and there fore would HAVE to have the same outcome. ANd like i said before, us sitting here now are part of the outcome of that reaction.
What does this have to do with predicting the future you ask?
Well, if you could create a computer program that models reality perfectly (which in itself would be an absolutely icnredible feat) then theoretically you could program it to simulate a reaction of the right amount of starting material it would simulate existence exactly as it is in the real world.
If you could then accellarate to the modern time within the program, you could be watching a computer simulation of yourself, watching a computer simulation of itself, watching a computer simulation of itself, ad infinty. Forward it more then...so that more reactions have occured in the program than in real existence. Because everything is caused bys omething else, caused by somtehin else caused by something else, etc, and because everything has a rpedictable outcome, you could see, say, what loto numbers would be drawn in tommorows loto. You could see who was going to develop a tumor and die, absolutel anything.
And of course you could look at the past, you could go to roman times and see if jesus existed, you could see how life began.
The applications of a program like this are limiteless...
Of course you are going to say "well if you can see the future in that world, isnt that changing the future in your world?
Well, remember in the virtual simulation there is a one of you lookin at the future to, so nothing is changed really.
Also, you could explore the universe with this machine, find if there is other life in the universe.
And of course with that program you could simulate any sort of experiment you wanted. You could create life inside a computer, play god, evolve it any way you wanted to.
Itd be a hell of a lot of fun...
But anyways
What do you all think of my dream idea?


 

another_someone

  • Guest
Re: A computer simulation that ccan predict the future
« Reply #1 on: 18/02/2006 03:30:47 »
This is the old Newtonian clockmaker, but the quantum guys would disagree with the first premise.

The quantum guys will say you cannot, even at the most theoretical level, describe the universe with absolute accuracy, only with probabilities of a given event happening.

From an information point of view, I would have a more basic problem.

Even if the universe was so deterministic that you could predict, with absolute precision, every event in the universe; you need to store the information about the universe.  If one were to create a very compact computer, the absolute minimum amount of storage one would need to store all the information about a single atom, would have to be another atom.  So it follows, that the absolute minimum amount of storage one would need to store all the information about a whole universe is another whole universe.

George

 

Offline DoctorBeaver

  • Naked Science Forum GOD!
  • *******
  • Posts: 12656
  • Thanked: 3 times
  • A stitch in time would have confused Einstein.
    • View Profile
Re: A computer simulation that ccan predict the future
« Reply #2 on: 18/02/2006 19:42:05 »
another-someone - I'm not sure another universe would be required to hold the information. If it was necessary to map every single particle, that would be the case. But would that be necessary? Very complex systems can be modelled already without needing to store every minute detail. Take the met office computer, for instance. It doesn't need to store information on every sub-atomic particle in the atmosphere to be able to predict with a fair degree of accuracy what the weather will be like for the next few days. OK, it can't say the edge of a rainstorm will pass over 93 Acacia Avenue, S****horpe; but a close approximation is often all that's required. Similarly, we can model the evolution of a star without a computer the size of a star.
 

Offline neilep

  • Withdrawnmist
  • Naked Science Forum GOD!
  • *******
  • Posts: 20602
  • Thanked: 8 times
    • View Profile
Re: A computer simulation that ccan predict the future
« Reply #3 on: 18/02/2006 20:47:11 »
Are you absolutely sure with no shadow of a doubt that by using the exact same starting conditions that an identical universe will manifest ?

I just hesitate to agree that all the consequences of this simulated identical universe will also by consequence follow the exact same path as ours.....I am sure there are some parameters somewhere that are open to differing repercussions despite the original premise.


Absolutely fascinating idea though....I think it's a great dream idea !

Men are the same as women.... just inside out !!
 

another_someone

  • Guest
Re: A computer simulation that ccan predict the future
« Reply #4 on: 18/02/2006 21:59:37 »
quote:
Originally posted by DoctorBeaver

another-someone - I'm not sure another universe would be required to hold the information. If it was necessary to map every single particle, that would be the case. But would that be necessary? Very complex systems can be modelled already without needing to store every minute detail. Take the met office computer, for instance. It doesn't need to store information on every sub-atomic particle in the atmosphere to be able to predict with a fair degree of accuracy what the weather will be like for the next few days. OK, it can't say the edge of a rainstorm will pass over 93 Acacia Avenue, S****horpe; but a close approximation is often all that's required. Similarly, we can model the evolution of a star without a computer the size of a star.



But I think you have made the key observation, what the met office are doing is merely an approximation, and that approximation gets ever more uncertain as one goes even a small distance into the future.

In a sense, what they have done amounts to a massively lossy compression of the information about the weather, and have used that lossy compression in order to recreate an approximation of reality.

Ofcourse, different aspects of the universe will tolerate different levels of compression while still retain high levels of accuracy, but never 100% accuracy, and so in the end the prediction will always fall apart some time in the future.



George
« Last Edit: 18/02/2006 22:05:47 by another_someone »
 

Offline DoctorBeaver

  • Naked Science Forum GOD!
  • *******
  • Posts: 12656
  • Thanked: 3 times
  • A stitch in time would have confused Einstein.
    • View Profile
Re: A computer simulation that ccan predict the future
« Reply #5 on: 19/02/2006 00:39:23 »
another_someone - I agree about the accuracy reducing as the forecasts look further into the future. But the point I was making is still valid. You don't need to map every single particle to predict what will happen. If someone invented a new alloy, you wouldn't need a computer with the capacity of all the atoms in a bar of it to predict how it will behave in certain conditions.
 

another_someone

  • Guest
Re: A computer simulation that ccan predict the future
« Reply #6 on: 19/02/2006 06:22:08 »
quote:
Originally posted by DoctorBeaver

another_someone - I agree about the accuracy reducing as the forecasts look further into the future. But the point I was making is still valid. You don't need to map every single particle to predict what will happen. If someone invented a new alloy, you wouldn't need a computer with the capacity of all the atoms in a bar of it to predict how it will behave in certain conditions.



It depends upon how precise you wish your prediction to be.

If you wish to understand that there is a 0.001% probability of failure of the material under certain conditions, then you do not need to model every atom in the alloy.

On the other hand, if you want to know that the metal part will fail at exactly 3:42 on the 3rd September 2008, and know exactly how it will fracture, even knowing exactly how many atoms will be in each part of the fractured material, then you will need to model every atom in the alloyed part, and every atom and cosmic particle, etc. that will come in contact with the metal part at any time until 3:42 on  the 3rd September 2008.



George
« Last Edit: 19/02/2006 06:33:28 by another_someone »
 

Offline DoctorBeaver

  • Naked Science Forum GOD!
  • *******
  • Posts: 12656
  • Thanked: 3 times
  • A stitch in time would have confused Einstein.
    • View Profile
Re: A computer simulation that ccan predict the future
« Reply #7 on: 19/02/2006 07:35:50 »
Couldn't it wait until 3:45 - halftime? :D

I take your point, but is that level of accuracy needed to predict the future? I suppose it depends on the scale at which you want to predict. To predict the behaviour of a system, you only need to model the pertinent components; but, yes, if you want to predict the behaviour of every single particle within that system, you would need to map every single particle.
I would contend that to all intents & purposes the behaviour of each & every particle would be immaterial in the overall scheme of things. If, all things being equal, the same initial conditions would result in the same outcome, would it really be necessary to say that life will appear on the next Earth exactly 4,532,684.598623754 years (or whatever) after its creation? Would it really matter if it was rounded to the nearest 100,000 years or so? The prediction would still be valid.

Incidentally, I used to design maintenance management software & predicting component failures was a vital aspect of the systems.
 

another_someone

  • Guest
Re: A computer simulation that ccan predict the future
« Reply #8 on: 19/02/2006 07:57:23 »
quote:
Originally posted by DoctorBeaver

I take your point, but is that level of accuracy needed to predict the future? I suppose it depends on the scale at which you want to predict. To predict the behaviour of a system, you only need to model the pertinent components; but, yes, if you want to predict the behaviour of every single particle within that system, you would need to map every single particle.
I would contend that to all intents & purposes the behaviour of each & every particle would be immaterial in the overall scheme of things. If, all things being equal, the same initial conditions would result in the same outcome, would it really be necessary to say that life will appear on the next Earth exactly 4,532,684.598623754 years (or whatever) after its creation? Would it really matter if it was rounded to the nearest 100,000 years or so? The prediction would still be valid.




It is not merely the level of accuracy, but even the level of certainty.

That there is a 0.001% probability that a component will fail within the next year is not the same as saying that one knows with certainty that it will or will not fail within the next year.  Something with a 0.001% chance of failure within a year, may fail within 10 seconds, or may survive in tact for a million years all you can say is what the uncertainty of a situation is, not predict as absolutely certain that a particular outcome will come about.

If one views the art of prediction as being about knowing the future with certainty (even an approximate future, but with absolute certainty), then one cannot rely merely upon probabilistic outcomes.



George
 

Offline DoctorBeaver

  • Naked Science Forum GOD!
  • *******
  • Posts: 12656
  • Thanked: 3 times
  • A stitch in time would have confused Einstein.
    • View Profile
Re: A computer simulation that ccan predict the future
« Reply #9 on: 19/02/2006 08:11:30 »
Looking back at Realmswalker's original post, I see I've drifted off course. He was indeed talking about precise events such as Lotto numbers. In order to do that, I totally agree with you that it would indeed be necessary to map everything.
 

another_someone

  • Guest
Re: A computer simulation that ccan predict the future
« Reply #10 on: 19/02/2006 11:38:02 »
quote:
Originally posted by neilep

Are you absolutely sure with no shadow of a doubt that by using the exact same starting conditions that an identical universe will manifest ?

I just hesitate to agree that all the consequences of this simulated identical universe will also by consequence follow the exact same path as ours.....I am sure there are some parameters somewhere that are open to differing repercussions despite the original premise.


Absolutely fascinating idea though....I think it's a great dream idea !




This is where I believe the quantum guys would go along with you, Neil.

Quantum theory says we cannot determine things with absolute precision, and the current interpretation does not merely say that we cannot measure it with absolute precision, but that there is innate randomness in its very existence, and thus even with the absolute knowledge of the entire universe (not possible according to Heisenberg), we still cannot predict exactly where the universe will be in the next instant.

Are the quantum physicist wrong?  Einstein certainly would be happy to hear they were, because he really did not like the innate randomness of quantum physics.



George
« Last Edit: 19/02/2006 11:39:33 by another_someone »
 

The Naked Scientists Forum

Re: A computer simulation that ccan predict the future
« Reply #10 on: 19/02/2006 11:38:02 »

 

SMF 2.0.10 | SMF © 2015, Simple Machines
SMFAds for Free Forums