Re: Memory Merging Possible For Close Duplicates

From: Lee Corbin (lcorbin@rawbw.com)
Date: Sun Mar 16 2008 - 13:55:48 MDT


Mike writes

> [Lee wrote]
>>
>> Few people would or should make that claim. But then, what you
>> are trying to explain is what the "nuances" really amount to.
>
> I am suggesting that their value is negligible and probably do not
> need to be preserved.

What nuances? Do I have to go back and find whatever post it was and
begin looking? Could you try harder to give context? Thanks.

>> > electromagnetic storage constraints). However, the bandwidth
>> > bringing data into your 'brain' can exceed the bandwidth to
>> > [memory storage][?]
>>
>> At least. And though we are just discussing principle, there
>> are the other senses that should be given passing mention.
>
> I assumed bandwidth covers the sum of all senses.

Yes. Right. Good.

>> second re-reading, I get it. That is, for example, do I squint so
>> that the incoming bandwidth isn't any greater than the bandwidth
>> to my memory? (There's another sentence that could have been
>> restated and embroidered with "That is...", or "In other words..."
>> ...
>> or trying to skim written material IN ORDER TO notice
>> less of it----ah! Thanks! Yes I do. Often, especially with pages
>> I already read or which seem overly redundant, I have found,
>> as per your prediction, that I can more accurately "remember
>> what I perceive".
>
> This goes to my point about how much memory is required or is even
> considered useful.
>
>> But still, if I take your question above quite literally, the answer
>> still is no, I think. BTW, what has this to do with Memory
>> Merging Possible for Close Duplicates?
>
> I agree that the answer should be no. Along the same intention of
> maximizing runtime should be maximizing resolution and bandwidth.
> There is also a point to be made here for the idea that intelligent
> compression allows a larger volume of data to be written to memory
> through the fixed-bandwidth channel. That's a different topic though,
> and you're already asking what this line of thinking has to do with
> the subject of close duplicates.

The compression you speak of could be done as you say, before
it reaches memory (in our highly oversimplified language and model
we're using here), or maybe memory itself that does it (e.g. at a future
time).

> What determines how close a duplicate is to the original? Whether the
> TE uses teleportation or computronium uploads, the principle measure
> of identity is ... what exactly?

It is similarity of structure, ultimately. Of course, we have various
measures of that (but we also have various yard-sticks that differ
one from the other a little, and even meter-sticks, but we firmly
believe in the utility of the 'length' concept, despite the dance of
atoms going on in physical objects).

Suppose that by some *obvious* [1] isomorphism, today Lee is
one bit string

11011011001110010110111111001010110100000011...
and tomorrow I am
11011011001110010110011111001010110100000010...

where two 0's happen to be changed to 1's. By all measures
of similarity, those two strings are still very similar. (It would be
easy to mention some particular measures used by mathematicians,
but the ones I'm completely familiar with without having to go look
them up aren't very applicable here.)

If the teleporter or copying machinery is working well, there will
be very few bit errors. I claim that the reality really is [Jef will
have a fit with that] that similarity of structure exists 100% for
all truly identical things (e.g. different carbon atoms in the same
QM state), and falls off directly as one thing starts less and less
to resemble another. In other words, resemblance does *not*
lie strictly in the eye of the beholder.

> It seems that often it is memories that are used to determine
> identity up to the point of copy, and that instances diverge
> from each other because of their new memories.

Yes. Or new experiences which create or destroy small parts
of the structure.

> I think the "what is identity" question is endemic to this "Close
> duplicates" discussion but it quickly turns into a distracting
> quagmire from which there is no return. It seems there are others
> here who believe memory is not the sole measure of identity, which
> gives rise to a completely different view of what a close duplicate
> would mean.

Yeah, you're right. Naturally, I think my views 100% sensible,
the views of many other worthies in the high 90's, and those who
believe that identity has nothing to do with memory in the low 10's.

> I ask about the relationship between experience of an event and memory
> of the event because it is relevant to how memory can be merged
> without losing the ability to reconstruct the experience. I would
> have to start a separate thread to discuss what it means to believe in
> a memory as experience, which (like the identity question) is relevant
> but distracting.

I'll do it! Thanks for the suggestion.

> I argue that it should not cause cognitive dissonance to say, "I went
> to the movies" at the same time "I went to the bookstore" (to use your
> original analogy) because both copies have a belief in their identity
> as Me.

Well, I don't quite agree with your "because both copies have a belief
in their identity as Me". To me, it's a fact of the matter, not a conviction.
Some people think they're Napoleon, but they're wrong. Yet it would
be possible for a fantastic coincidence of molecular motions to produce
the actual Napoleon. (He might have a hard time proving that, at least
until he talked to enough historians and linguists!)

> I also currently believe that each copy, for the duration that
> each existed, was Me.

So do I.

> As long as I am not bound by a belief that my identity can only exist
> as a linear series of experience, there is no problem with being several
> simultaneous duplicates.

Well, that's two of us. Actually, there may be three or four here who agree
with us! :-)

> Is that a clearer statement of support for my original assessment that
> close duplicates should merge at least as easily as they are created?

I don't see how that last statement follows from anything above.
We may be having a terminological problem here. In principle,
nothing is easier than creating duplicates whose subsequent
experiences and memories diverge. In fact, they can merge
away from each other arbitrarily far, and become---to my way
of thinking---different people entirely.

But how in the world to merge them? What about temporal
conflicts, like "Oh, man, am I confused! I could swear I went
to the bookstore last night until it closed at 11pm. But I can
also swear I was watching "The Incredibles" from 9pm to
11pm." Somehow we have to reconcile this in his brain.
And that's just an easy example. Much more difficult conflicts
can be described. So merging is vastly more difficult than
copy generation.

Lee

[1] A detailed bas-relief of the Mississippi Valley is "isomorphic"
to the real Mississippi Valley, where the scare quoted term doesn't
have exactly the same meaning that it does in mathematics. Here
I'm using the term much as the way a mathematician might use
"homomorphism". As for it being "obvious", if people all died and
an alien came along, he *would* be able to stop the homomorphism.
He'd say "Hmm, those entities were making a map of that valley
for some reason."

Likewise, there is an easy "isomorphism" between the sentence
"now is the time for all good men to..." and the string

1415220009190020080500200913050006151800011212
000715150400130514002015...

which people who like codes and ciphers will have no trouble
decrypting. The same is true of a bit string representing a person.
Sooner or later a highly intelligent entity will spot whatever coding
scheme has been used to encode the person as bits.



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:02 MDT