Re: Everett

Hal Finney (hal@rain.org)
Wed, 30 Jul 1997 18:54:45 -0700


Brent -

> Do you have any kind of references to where I
> could read more about these "many studies showing how this kind of
> environmental interaction transforms the state from a quantum
> superposition to a classical mixture"? Maybe even a source
> approachable by someone as stupid about physics as I am?

I don't have any non-technical references handy. One on-line paper
I saw recently was http://www.sns.ias.edu/~max/collapse.html, which
I believe discusses some related phenomena. However it is very
technical.

> So, then, is a quantum computer merely something that gets
> quantum physical phenomenon to represent and logically operate on
> information like a regular computer does with "classical" hardware?
> If so, then why couldn't this "superposition" information simply be
> represented by a sufficiently complex and appropriately programmed
> "classical" machine? Maybe it would be easier with "quantum"
> machinery but why couldn't it be done with classical hardware?

It depends on what you want to do. If all you want is to do calculations,
as with most of the recent interest in quantum computers, then you're
right. Any calculation doable on the quantum computer can be done on the
classical computer. The interest is that theoretically a quantum computer
can do some calculations much faster than a classical computer.

However, for the purposes of the thought experiment John Clark described,
invented by David Deutsch (and predating the recent interest in quantum
computers, although Deutsch pioneered the field in the 80s), we want to
do something else. We want to measure the state of a quantum system, and
bring that measurement into the quantum computer, preserving its quantum
nature.

Consider a photon, or some other particle. It enters a computer's sensor,
and the computer will then work on the information obtained from that
photon.

In a classical computer, we will basically absorb the photon, and measure
some of its state. However, because of the Heisenberg uncertainty
principle, we will inevitably measure only a part of the state. Something
is always lost.

In a quantum computer, we must (somehow) accept the photon into the
computer in such a way that the wholeness of the quantum state is
preserved. It may not stay in the particle, it may be transferred to
some other particles, which represent and preserve the original state.
Then, very carefully, we must operate on this representation, always
protecting it from noise and preserving the coherence information.
Finally, in Deutsch's example, we may eventually run some part of the
calculation backwards and actually re-create the photon as it was when
it entered the computer.

This is very demanding. Today we can bring a photon in and put it into
a loop, with amplifiers, and preserve most of its state for some fleeting
instants. To actually do a lot of calculations on the photon (enough to
simulate a conscious mind!) while preserving the phase and state information
so perfectly will be quite a feat, to say the least!

Hope this sheds some light for you -

Hal