[ExI] Is a copy of you really you?

[Keith Henson]

On Sun, May 31, 2020 at 7:43 AM Robin D Hanson <rhanson at gmu.edu> wrote:
>
> On May 30, 2020, at 6:17 PM, Keith Henson via extropy-chat <extropy-chat at lists.extropy.org> wrote:
>
> Robin D Hanson <rhanson at gmu.edu> wrote:
>
> (re my objection to copies)
>
> I disagree; I wrote a whole book describing a reasonable world where many copies are made:
>
>> And you might be right, nobody has a lock on what the future will turn
>> out to be, certainly not me.
>
>> However, I have two objections.  My work in evolutionary psychology
>> leads me to feel very uncomfortable about using humans as a model for
>> building minds.  Humans have traits like capture-bonding and those
>> which lead up to wars that were selected. Having one of those (or ones
>> we don't know about) activate in a powerful AI seems intolerably
>> dangerous.
>
>
> That same argument would suggest trying to eliminate the human minds in meat brains, as well as those in artificial brains.

We tolerate meat brains (no choice actually) because most individual
humans have relatively litle power to mess up the world.  What
concerns me is giving a human mind copy vastly more power in an
artificial brain.

Humans seem to have evolved psychological mechanisms that detect
"looming privation," resource shortages, and for good evolutionary
reasons respond by attacking neighbors and taking their resources.
(After a perious to build up xenophobic memes.)

Consider a raw human mind mapped into a powerful AI with substantial
enhancement.  Would such a mind have increased concern about a future
resource crisis?  I suspect that might be the case.  Would this
concern turn on the evolved psychological mechanisms and lead to an
irrational path to war?  I don't know.  However, I would suggest
caution.

> But if you see human brains as holding most of the value in the world, you’ll want more of them.

Within limits, yes.  Falling income per capita seems to flip the
behavior switch into war mode, while rising income per capita seems to
turn it off.  Rising income per capita seems to be what caused the IRA
to go out of business.

Optimum human numbers might be a lot higher than the current,
especially if we spread into space.  But I suspect that very large
numbers where the resources per person were sharply limited would not
be a good idea.

>> The other objection is right out of economics.  If it is cheap to copy
>> something its value falls to the marginal cost.  I.e., one Michael
>> Jordan is valuable, at least to him.  10,000 of them would have an
>> interesting time finding teams to play with.
>
>
> The fact of diminishing marginal value is not at all an economic argument against larger quantities of anything.

I don't understand economics, but this feels wrong or at least
incomplete.  If you have a poker chip and there is a way to make them
cheaply by the thousands, well, you are not hurt and may be better off
if you have use for lots of poker chips.

But if you, yourself, are duplicated by the thousands, millions, or
billions and there is some limit on the number of poker chips that can
be made, then a large numbers of copies of you are going to make you
poor in terms of chips.

It seems to me that there is a fundamental difference between things
that can be owned and things (people?, ems?) who owns stuff.  If an
ems' mental organization is copied from humans, are they more like
people or poker chips?

Keith

PS  Jack Vance's story "The Last Castle" describes the revolt of the
"Mechs," an alien race that has some similarites to ems.




> But, as I freely admit, you might be right and a vast number of copies
> could be the path to utopia.
>
>
> I’m not claiming utopia; I’m not sure the concept is even coherent.
>
> Robin Hanson rhanson at gmu.edu
> Future of Humanity Inst., Oxford University
> Assoc. Prof. Economics, George Mason University
> See my books: http://ageofem.com http://elephantinthebrain.com
>
>
>
>
>
>
>
>
>