X-Message-Number: 8312
From: "Peter C. McCluskey" <>
Date: Wed, 11 Jun 1997 22:56:34 -0700
Subject: Bits or Souls?
References: <>

>Message #8300
>Date: Mon, 09 Jun 1997 09:09:46 -0700
>From: Peter Merel <>
>Peter C. McCluskey writes,
>> It is fairly easy to write a simulation such that no behavior of the
>>simulated entities can crash the simulation -
>That may be, but it's not possible to prove - cf. the General Halting 
>Problem in any Computability textbook. I quite agree with you that

 It says undecidable programs exist, not that all programs are undecidable.

>our Bob is off the beam with this "there would be bugs ..." bit, but I 
>think that the disagreement here is philosophical rather than logical.
>To we folk who regard information as a physical property, Bob's position 
>looks irrational, but to folk who regard identity as a physical property,
>the info position looks irrational. Interesting, ain't it? Maybe we should 
>investigate the ascription of physicality a little more, rather than 
>keep wasting bits on slanging ...

 Yes, it would be interesting if we could pin them down to a scientific
test of identity.

>Message #8302
>Date: Mon, 9 Jun 1997 15:11:54 -0400 (EDT)
>Subject: #8296

>In any case--again, according to the terms of the discussion as I understand
>them--the simulated world would be as much like the real one as the
>programmers' knowledge permitted. Hence, under any reasonable assumptions,
>surprises (discoveries) would be expected in a real world, not in a simulated

 Non sequitur. I expect that if humans a century from now try to write
such a simulation, the results would contain about as many surprises as
we see.

>2. I said that programmers usually write bugs in complex programs, and
>simulated programmers in sims and subsims would also. Mr. McCluskey says (a)
>you can tend to avoid bugs by keeping the rules simple, say quantum
>mechanics, and start with the big bang; and (b) if there is a bug at sim
>level X, you could start over at some point and eventually get rid of bugs,
>and a typical simulated person would observe no evidence of crashes.
>(a) Again, this changes the rules of the discussion, as I understood them. We
>are not starting a new universe; we are trying to continue our present world
>and present lives by simulation.  The rules cannot be kept simple, because
>the world and the people in it are complex.

 I guess I'm confused about what you are trying to accomplish with your
argument. If you narrow the set of simulations enough, you can certainly
demonstrate that we aren't living in a simulation belonging to that set.
I am trying to demonstrate that:
 a) we might be living in a simulation of some sort (but not that we
   might recently have been switched into a simulation without knowing it).
 b) you can't show that being uploaded into a simulation would cause
  problems that we all agree are serious.

 I deny that 22nd century humans will need to write simulations complex
enough to be unavoidably buggy in order to upload. A simple approach
that might work is to find a good model of how atoms behave, analyze
a person to atomic precision, and run the results of the analysis.
(Not that I expect people would want to upload will be paranoid enough
about bugs to pay for such an expensive method).

>(b) First, if you have to backtrack and rejigger the program, then my
>proposition was correct--the simulated lives (or at least some of them) did
>not continue as the original would. Second, the bug might (it seems to me)
>not necessarily "crash" the program, but instead just cause goofy results.

 Which wouldn't necessarily cause any harm. The simulator could insure that
the simulated beings wouldn't be aware of being restarted. Goofy results
could either be reason for the simulator to restart the simulation from
a recent backup or could constitute the surprises that you think are unlikely
in simulations.
Peter McCluskey |                        | "Don't blame me. I voted
 | http://www.rahul.net/pcm | for Kodos." - Homer Simpson
 | http://www.quote.com     | 

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=8312