So apparently I have been missing some of the most interesting arguments on Earth.(1) Basically, those made by Nick Bostrom, who I had not heard of a few hours ago. I am more than a little pissed about this but lets leave that alone.
Bostrom argues that the there are three possibilities.
1) Most human like civilizations will never reach the point where they can simulate reality
2) Most civilizations who do reach that stage will not want to create simulations
3) We are probably living in a simulation.
The reasoning for this seems obvious to me, so I won’t go into it. You can find it at http://www.simulation-argument.com if you want. (HT: TGGP)
When I first heard a sketch of the argument in passing, I heard it phrased that (3) was likely. My response was what about (2)!
However, after having read Nick Bostrom’s paper (2) seems unlikely to me.
Why such a quick turn around?
Well, if a civilization becomes so advanced that it can create simulations then it is highly unlikely that it has not thought of the simulation argument. If it has thought of the simulation argument then, presuming the people are anything like us, they will want create the simulations as a test of the simulation argument.
That is, the desire to test (1) makes (2) unlikely.
Now how likely is that such people will want to test (1)?
This is admittedly very rough but it seems unlikely to me because such a civilization has already progressed to the level of being able to create a simulation. This means that they have extensive knowledge of the workings of the universe. Currently, I cannot imagine how this could be done without some interest in the scientific process. That is, the desire to test hypotheses.
Now, of course this is a statement about my imagination, not reality. However, it still seems that my subjective estimate of the probability of (2) should be low unless I encounter some evidence that makes it seem more feasible that a sufficiently advanced civilization will not want to test (1).
So as of right now it seems that either (1) or (3) is most likely.
(1) Of course this comes with caveat, that I now know of. Since it is obviously, possible, indeed likely, that I am missing out on even more interesting arguments. Do not hesitate to post them if you know them.
2 comments
Comments feed for this article
Monday ~ February 15th, 2010 at 5:54 pm
Rick Russell
If we are simulants, we can only speculate about the universe that the “basement species” lives in, but the universe we live in seems to have intrinsic randomness at the quantum level.
Right now, it would not actually be possible to simulate the evolution of the universe with any amount of computing power, as the physical events that make up the universe include a past which cannot be truly known and a future which cannot be predicted, due to uncertainty at the quantum level.
He even alludes to one aspect of the problem in his paper:
“it is likely that the human central nervous system has a high degree of redundancy on the mircoscale to compensate for the unreliability and noisiness of its neuronal components”
“Noise”, by definition, cannot be reproduced or predicted. If such noise exists, then creating a system where it does not exist, or dealing with it is not required, is actually simulating something quite different.
By the way, you should look up the science fiction stories of Greg Egan. Virtually all of Egan’s short stories deal with this problem in some capacity.
Tuesday ~ February 16th, 2010 at 12:14 am
Thorfinn
The best part about this argument that if we test 1), and find that we can indeed generate simulations, the odds become very likely that 3) is in fact the case. So this is in fact testable science!