[extropy-chat] Timeshifting

Adrian Tymes wingcat at pacbell.net
Thu Feb 26 20:07:18 UTC 2004


--- "Robert J. Bradbury" <bradbury at aeiveos.com> wrote:
> Harvey (and Adrian) I understand this.

Judging by your comments, it does not seem that you
understand the full implications.

> My best
> guess right now is that
> genetics would only get you 50-60% of an identity. 
> But that is a start.

No, it isn't.  Even if you replicate almost all the
external factors - genes and environment - you can
still wind up with someone who has differences those
close to the person would find significant, at least
emotionally if for no other reason.  And in practice,
there's always some little bits of information that
only a person knows, which will escape any attempt by
others to discover without said person's cooperation.

> I probably have most of my financial records in
> detail for about 20 years
> (so one could easily determine "what has Robert
> done?").

To an extent.  But it wouldn't say anything about that
which cost you no money but could have had a
profound impact upon your life.  For example, your
choice of electives in college (or high school, if
they offered such): this would have significantly
affected your knowledge base, but after 20 years,
there are probably no records remaining, save maybe
your own memory.

> I have (or the
> net could provide) many of my email posts dating
> back perhaps 20-25 years.
> Combine this with photographs of where I have been
> and what I have seen
> and done and one starts to come up with a Robert-B
> that knows much if
> not more than what Robert-A knows about what he has
> done.

True.  But it does not necessarily allow perfect, or
similar, duplication of how Robert-A thinks.

Also, consider: even if perfect duplication were
possible, people change over time.  Who's to say that
Robert-B would evolve in the same way Robert-A would
under the same circumstances?  This would seem to be a
requirement of perfect duplication, if people are
deterministic as you state.  Does the answer change if
Robert-A is still alive and experiencing the same
circumstances?  If so, why?

> Place this on top of a genetic profile that
> determines motivations,
> feelings, etc. and one starts to develop a
> personality that determines
> who you are, how you act, etc.  I would assert that
> that would be much
> better than a face recreation on top of an actor.

I would assert that an actor, given the same
information (perhaps an intelligence operative or the
like), could do far better than an AI simulation
today.  I would also assert that, by the time AI
progresses such that the above is no longer true, it
will likely be practical to create backups of a
living person's memory, which will be higher fidelity
emulations than a simluation based on external
factors.

> I
> have no problem
> with the fact that twins are different -- but a
> relatively detailed
> simulation of the end-point
> personalities/life-styles, etc. should
> reveal how/when/where there was a divergence (if
> enough information
> is available).

This is analogous to why transporters are not (yet)
practical, even though bits of light have been
transported: the amount of information you'd need to
create a good simulation exceeds what today's
supercomputers can handle.  It is questionable
whether, in fact, that information might not be
unobtainable due to the Hesienberg limit; we're
talking stuff like excitation levels which would
affect exactly when certain neurons fire, and see
chaos theory for proof that that makes the bigger
picture (the person) not 100% predictable.

> Doesn't anyone out there have any friends,
> significant others, etc.
> where you *know* "If I say this -- they will say
> that?".

If you want personal experiences - even my closest
friends, I can only say that about a certain limited
range of topics, and I know I will not be 100% correct
in my predictions.

> That seems
> to suggest that human beings behave in a relatively
> deterministic
> fashion.  That would imply you can backtrack
> development processes.

There is a *MAJOR* difference between 100%
deteministic and 99% deterministic, and you've just
hinted at it.  100%, like a computer, you can
sometimes backtrack (assuming only one path of inputs
and starting conditions would lead to the observed
output).  99%, you can't, unless you're willing to
accept the possibility that you could be wrong - and
those possibilities multiply when you consider
multiple events.

Consider: what is 100% to the 1000th power?  100%.
But what is 99% to the 1000th power?  Somewhere below
.05%.  And the number of potential inaccuracies in
this situation is likely far higher than 1000.  (How
many seconds have you been alive?  Consider that
neurons fire more than once a second, and each
firing is a discrete "event"...and that's not taking
into account the genetics and pre-natal environment
that set up your brain in the first place.)

> (Just as one futuretracks moves when playing chess.)
>  I would
> suggest if humans did not behave in a relatively
> deterministic
> fashion that society could not have developed to the
> extent that
> it has.

Nope.  Society still works even with high but less
than 100% determinism.  It only has to be likely, not
guaranteed, that people will act in certain ways.  The
exceptions provide mutations to social memes; the
"mutation" rate is low enough that bad memes can
usually be trounced, while good memes will usually
thrive.  (Of course, it's still high enough that
people have to take precautions against criminals,
terrorists, and so forth...but that's proof that it
exists to some significant degree.)

(Pause for reflection on a semi-unrelated tangent.
The core problem with ideas like this is that they are
easy to disprove - except to those who believe in them
beyond reason.  They then try to implement them,
inevitably to great personal distress when the failure
becomes undeniable yet they try to deny the failure.
This is part of the "logic" behind luddites and the
like, except that they have a grossly limited view of
what's possible; they try to claim as impossible the
things we go ahead and implement, proving them and
their fears wrong.  Who's to say what's possible and
what's not possible?  Who's to say that, for example,
people are or are not in fact deterministic - and
would it be possible and practical to measure enough
information to discover this program that a person
follows?  The only way to tell is by debate like the
above.)



More information about the extropy-chat mailing list