From: hal@finney.org
Date: Fri Jul 14 2000 - 10:34:47 MDT
I missed who Damien was quoting here:
>I usually try to avoid strong language, but the phrase "total
>bullshit" comes to mind. Deterministic is deterministic. It's a
>computer. Everything inside the box is deterministic. It doesn't
>matter how complex their random-number algorithm is; unless they
>actually have a radioactive isotope in there, it's deterministic.
Actually, radioactivity is not necessary. There are other
nondeterministic sources of noise in nature. Thermal noise can be
detected via a reverse biased diode or similar components. Thermal noise
is deterministic in a Newtonian universe, but in practice the molecular
motions that cause the noise will be affected by the uncertainty principle
and so will partake of quantum randomness.
Some newer computers include a component from Intel which is a true random
number generator chip based on thermal noise. It is part of their bus
control chip. I'm not sure exactly how this RNG is used by Windows in the
installed state, but it would certainly be possible to have a program that
would use bits from this chip as true random bits for psi experiments.
Damien wrote:
> I take it what's being said here is this: even if the sloppiness of Windows
> makes the sampled data *unpredictable* to a human, nonetheless they occur
> in a deterministic sequence that is set in ineluctable motion from the
> moment the seed enters the algorithm. This would be good enough for a
> precognition experiment, say, but not for one of mental *influence*.
Interestingly, this would suggest that it would be possible for a
psychic to distinguish a true-random source from a deterministic,
pseudo-random one. This is thought to be impossible if certain number
theory problems are hard, like factoring. Of course, factoring would
not be hard for psychics; they could just guess the factors a bit at
a time, using some of the noise-reducing techniques. They could also
break ciphers by guessing keys.
This is the kind of test which I'd like to see: generate a crypto key,
encrypt a message, then destroy the key. See if the psychics can
guess the decryption key and decrypt the message. There'd still be a
possibility of cheating, but if the generation and destruction were done
in some kind of public manner, psychics all over the world could spend
as much time as necessary to get the answer.
Hal
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:29:56 MST