# RE: Real vs. Simulated Ecologies [was: Moravec article available]

From: Billy Brown (bbrown@transcient.com)
Date: Wed May 03 2000 - 15:46:13 MDT

hal@finney.org wrote:
> Billy writes:
> > The thing you have to remember about a VR is that it isn't an alternate
> > universe. Any transformation that happens in the virtual world has to be
> > implemented as an actual transformation on real particles in a physical
> > system. It doesn't have to be the *same* transformation, but it has to
be
> > *some* transformation of equal or greater information content. No matter
how
> > cleverly you design the VR, it will always have this constraint due to
the
> > simple fact that the virtual world has no independent existence.
>
> So you are in the VR, and you have built a computer which holds a 1
> or a 0, and you want to erase it. Obviously there is no need for the
> VR to simulate anything becoming hotter. It can simulate any laws of
> physics it wants to, and it need not be constrained by the usual laws
> of thermodynamics.
>
> However in the physical world if the VR erasure is mirrored by a physical
> erasure, that must dissipate heat. The amount dissipated need not have
> any particular bearing on the amount that the VR person would calculate,
> though.

There is no "if" about it. In order to erase a virtual bit in the VR, there
must be *at least* one real bit being erased in the real world (in fact,
unless you are extremely clever about how you encode the VR's data there
will be lots and lots of real bits for every "virtual" bit). So, the VR
person might not be able to see the energy dissipation, but it still happens
and it still represents a limit on the amount of computation the VR person
can do.

> For one thing, the amount of heat dissipated per erasure depends on the
> temperature. The simulated temperature in the VR doesn't have anything
> to do with the temperature of the physical system. The VR may appear
> to be erasing a bit in a room temperature computer while the physical
> computer running the VR is cryogenically cooled.

True, but that has no effect on the amount of computation the system can do.
The constraint is the performance of the real, physical computer running the
VR. No matter how fancy the VR is, no VR person (or computer) can ever do
more computations than the underlying hardware.

> It could also be possible to reduce the cost of erasure or even avoid
> it altogether, by the trick of running the simulation forward while
> preserving all the extra bits, then running it backwards to erase the
> calculation and the extra bits. You can do this every so often and
> just snapshot the state that is needed to proceed with the calculation.
> The result is that many erasures wouldn't get counted, only those which
> are necessary to go from one snapshot to the next.

You don't need a VR to do that. In fact, it would be much more efficient to
dispense with the VR entirely.

A VR is very useful for modifying the subjective experiences of people (or
other information processing entities), but it doesn't increase the amount
of computation that can be done with a given supply of physical matter,
energy and time.

This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:10:26 MDT