Re: Fury over bishop's comments

Peter Ashby wrote:

Gareth McCaughan <Gareth.McCaughan@xxxxxxxxx> wrote:

Peter Ashby wrote:

So given that we have this ability, why would we not in principle be
unable to detect information flow from our region of space to the
'hardware' and back again.

This is where you lose me; it seems like a total non sequitur.
By what sort of means do you imagine that we might detect this?
It would be, quite literally, outside our universe.

Only once it had left it, but the leaving would leave a negativity, a
sink in this one. The return would cause a positivity, a source in this
one. As I say I think this would of necessity be asynchronous making it
even easier to detect.

And here again you lose me. What sort of negativity or positivity?
What does that *mean* when you're talking about information? Information
can be copied without loss. When I copy a file from one computer to
another it doesn't leave a "negativity" on the computer it was copied
from.

No, but it does produce a positivity of information on the duplicate
disc which is easily detected.

By people or things with access to that duplicate disc.

Imagine that the copy is made by attaching extra wires
to the connectors between the original disc and the rest
of the computer, and watching the bits fly past. (It might
take a while to find out everything that's on the disc,
but even a partial copy could be very useful.) And let's
suppose it's done very carefully so as not to disturb
any of the signals enough to change a 1-bit to a 0-bit
or vice versa. And let's suppose (as may or may not be
the case in reality) that the computer isn't equipped
with things like analogue-to-digital converters measuring
temperatures or magnetic fields or whatever, that might
be influenced by this wiretap.

Then there is, quite literally, no way for any software
on the computer to notice that the information is being
copied. As far as programs running on the computer are
concerned, the copying is perfectly undetectable.

That's perhaps a slight idealization. Presumably there
are occasional random errors of various sorts in the
computer hardware, and the presence of this wiretap
might make them more likely. So, we stick some error-correcting
codes into the protocol by which the disc controller talks
to the disc hardware. (Far from trivial with real existing
hardware, of course, but this is all a thought experiment.)
Then it's perfectly straightforward to reduce the effect of
occasional errors on those wires so far that (say) the
probability that they ever produce a change in any bit
read from or written to the disc is less than 10^-100.

It also uses local energy to duplicate
the information which is also detectable. Information production uses
energy and so produces patches of increased entropy. Ditto incoming
information which produces patches of decreased entropy. The exact form
of the encoding of the information is not relevant to this fact, it will
still be detectable at the level of entropy.

By something outside the computer, observing it. Not by software
on the computer. Our hypothetical universe-simulator is software.
It has no ability to sense these changes in energy and entropy
unless we provide hardware to let it do so. There's no reason
why we should.

In particular, if the software we're talking about is a
universe simulator then there is *no reason at all* why
changes in energy or entropy in the computer running the
software need correspond to changes in energy or entropy
inside the simulation. If I write a program to simulate
simple Newtonian mechanics and then put my computer in
a centrifuge, the objects in the simulation won't change
their dynamics. (Unless the centrifuging breaks the computer,
in which case the simulation will probably stop running
altogether.)

You still haven't explained how we would go about detecting this
alleged leakage of information if our world were in fact a simulation
in someone's computer.

See above and I have by reference to SETI, if SETI can do what it does
then the inhabitants of a simulation can do what I propose.

I still don't know what you're proposing. Once again: Suppose
Alice says our universe is running inside someone's computer
and Bob says it isn't. What experiment can Alice and Bob do
(or could they do, given sufficient resources) to provide
evidence about which of them is right?

BTW if you remember our episode of much heat and no light over Boltzman
Brains it turns out that nobody intended it to be taken seriously, it
was a reductio ad absurdum argument as a refutation. Detailed here:
<http://cosmicvariance.com/2008/01/14/boltzmanns-universe/>

As I said all along.

| 9. Some people think that any such theory has the consequence
| that anyone who believes the theory should conclude first
| that s/he is in fact a Boltzmann brain and second that, as it
| were, all cognitive bets are off (because in that situation
| their understanding of the universe is almost certainly broken),
| and accordingly that we should reject such theories.

So there :-).

--
Gareth McCaughan
..sig under construc
.

Relevant Pages

  • Re: Havent done anything real with OOP yet.
    ... The beam turns on and off on the order ... computed every 7/60th s in the Computer hardware element. ... >>observation from the Resistor, though the Cell, Thermometer, and ... In that situation you have a simulation error because the ...
    (comp.object)
  • Re: It seems that port mapping causes no skew
    ... variable assignment as a method to save a result to be used by ... defined semantics are for the mythical (simulation) aspect, ... trivially obvious to a real hardware guy, ... the signals get combined to belong to the same "net". ...
    (comp.lang.vhdl)
  • Re: Havent done anything real with OOP yet.
    ... >>will be easier to deal with and will reduce risk of simulation errors. ... >>Thermometer, heat transfer to the Cell, and heat transfer to the heat ... > thermometer, cell, heatsink, and etc., don't exist. ... A simulation models something, in this case hardware. ...
    (comp.object)
  • Re: Learning C?
    ... I've answered this before - using C++ objects to simulate hardware ... especially when the target is a small 8 or 16 embedded device ... What capacity for exploding in a SW simulation? ... better hardware emulators can also be run without the target HW and give ...
    (comp.lang.c)
  • Re: Havent done anything real with OOP yet.
    ... >OOA modeling can be used for hardware design. ... >superficially true enough to support hardware simulation, ... >until the last cell has completed its computation. ... sort of rigamarole. ...
    (comp.object)