 |
| |
 |
| |
|
| |
 |
The
term uploading, as used on the Net today, refers to taking
a computer file and sending it to a different computer,
usually a high-powered server of some kind. In futurist
circles, however, it also has a different meaning: the transferal
of human minds out of their flesh containers into stronger,
more flexible mechanical substrates.
It’s a strange idea, but also natural, in a way. After
all, are you your body, or are you your mind? If you lose
an arm, you’re still you; if you get a brain tumor
and have a chunk of brain cut out, you’re still you,
though perhaps a little different. But then when you die,
your brain stops distributing electricity throughout itself,
and you’re not there anymore. Why should your mind
be tied to the particular container it was born in? Your
mind is information: it’s not your brain, any more
than a computer file is the hard drive that it’s stored
on.
It sounds outlandish perhaps, this kind of uploading, but
what does it require technologically? Faster and more capacious
computers than we have today, and more accurate brain scanners.
Computers get better and brain scanners get more accurate
every year. It’s only a matter of time – perhaps
just decades – before you’ll sit down at your
computer in the morning and bring up windows featuring a
word processor, an action game, and uploaded Uncle Harry.
Of
all the future tech topics I’ve ever encountered,
none leads to deeper philosophical dilemmas than uploading.
Dilemmas related to cloning, AI and such fade into insignificance
by comparison. The seemingly simple question “Is my
upload me or not?” has enough different shades to
fill an art gallery.
One case of uploading is the creation of an upload with
a completely separate physical embodiment from the original
mind, while the original mind is left alive and intact.
There’s you on the computer, and you in your good
old human body, staring at each other. Richard Kennaway,
in a 1992 article summarizing the various wildly contradictory
views various theorists have put forth on this issue , made
the following statement about this situation: “One
of the few points of agreement. Everyone seems to agree
that in this case, the original and the duplicate are separate
persons. Neither experiences the thoughts, feelings, sensations,
etc. of the other.” But in fact, I’m not sure
that I agree with his statement; so I’m quite certain
it’s not true that everyone else besides me does.
There has been a lot of interest in supposed paranormal
phenomena associated with identical twins: sometimes, it
seems, they can sense each others’ mind-states automatically,
especially in cases where highly dangerous or otherwise
emotionally charged events are involved. If identical twins
can in this sense experience a consciousness-overlap, then
perhaps two identical minds will do so in a more extreme
way. Perhaps there will be a psychic link between a mind
and its corresponding upload. Quantum physics does not rule
this out: the brain may be a macroscopic quantum system,
and there may be bizarre quantum nonlocality effects joining
a mind to its digital copy. Of course, this is pure speculation,
but it is speculation that is consistent with all known
physical laws.
But what if your human brain is destroyed during the uploading
process? Then there’s only one you left – the
one in the computer. Is it you? Or are you dead, and a new
being with great resemblance to you newly created?.
A yet more difficult conceptual dilemma is posed by the
notion of gradual uploading. This notion refers to the temporary
hybridization of a brain with a computing device. The computing
device then gradually takes over more and more functions
of the brain, until eventually the computer is doing everything
and the brain is doing nothing, and the brain can safely
be allowed to die. Suppose this process is perfected –
then what happens, subjectively? There is a continuous consciousness
throughout the whole gradual uploading process. Is it not
then obvious that “you” are preserved through
the process? Or will there be more of a feeling of transition
from one consciousness into another? Will it feel like another
you gradually sucking the original you away, or will it
just feel like a natural change, similar to going through
puberty – you’re very different at the end,
but it’s still “you” all along?
As well as presenting new dilemmas, the concept of uploading
reminds us of the generally-ignored philosophical confusingness
of everyday life. In what sense, after all, am I the same
“me” as I was when I was 5 years old? Take a
case such as John Nash, the great mathematician whose plunge
into schizophrenia and subsequent recovery were portrayed
in the recent book and film A Beautiful Mind. After he became
schizophrenic, was he really the same John Nash or not?
Anyone who’s been married a long time is likely to
have observed their spouse change tremendously -- still
preserving some kind of essence throughout the changes,
but how much? Many divorces occur because one or another
partner has changed into an extremely different person.
As the late transhumanist theorist Sasha Chislenko put it
in a long e-mail on uploading, “if we just look at
our own lives, we go thru so many transitions that hardly
preserve our identity in any reasonable definition of this
word.…”
He ended the e-mail as follows:
Now I am ending the message being attracted by a personal
singularity
point that I try to cross daily and that is only an example
of a rich
variety of identities that constitute something that I [?]
call "myself".
It's 2 a.m., time to sleep. I am discontinuing my conscious
existence,
and leaving this message together with all my belongings
to the tomorrow's
being who I have never seen and who will probably think
that he is *me*.
Hope he'll mail it.
Ever
go to sleep thinking hard about something, tell yourself
emphatically not to forget about it overnight, and then
wake up 7 hours later completely unable to remember what
the thing was? In what sense is the morning biologically-embodied
you “the same” as the evening biologically-embodied
you?
And the puzzles of identity get even trickier when one considers
that many uploads won’t want to remain close copies
of their human minds of origin. Why bother, when there’s
a whole new assemblage of things to be? Why not add direct
mind-links into a calculator, a programming language compiler,
a paint program, a music synthesizer; why not link up to
global satellite networks as new sense organs? Why bother
sleeping? Why not create new states of mind between dream
and sleep? Why not freely edit out bad memories, why not
create simulated drug experiences as far beyond LSD as LSD
is beyond clove cigarettes? Why not simulate the sensation
of having sex using 19.7 sex organs at once? Why not fuse
your knowledge base with that of your wife, child or best
friend? And once you’ve done all these things, are
you still you or not?
I think these are fascinating issues, but I classify them
along with issues like “Is the world really there
or just an individual or collective mental projection?”,
and “Is anyone besides me really conscious, or are
they all just automata?” These are deep philosophical
dilemmas which none of us ever satisfactorily answers, but,
we go on living anyway. In the same sense, I believe, uploading
will happen, and will continue to be philosophically problematic
even after it becomes a commonplace of life. Life is philosophically
confusing, with and without the help of advanced technology.
Perhaps an upload will feel like it is and is not the same
person as its previous biological incarnation – but
it will go on living anyway, as we all do, in spite of the
numerous philosophical dilemmas that life poses.
On Principia Cybernetica (http://pespmc1.vub.ac.be/), one
of the jewels of the Net, a website devoted to futuristic
and system-theoretic topics, it is opined that “The
decline of traditional religions appealing to metaphysical
immortality threatens to degrade modern society. Cybernetic
immortality can take the place of metaphysical immortality
to provide the ultimate goals and values for the emerging
global civilization.” This is a bit optimistic in
my view: I’m not so sure that uploading and the ensuing
virtual immortality of the mind will in itself provide new
goals and values. But surely it will be a major part of
a new commonsense philosophy, as different from our contemporary
perspective as the latter is from the Stone Age worldview
of African pygmy tribes.
Fascinatingly,
the philosophy of uploading came up in a discussion between
systems theorist Franciso Varela and the Dalai Lama, a few
years back. Varela is a Buddhist as well as a biologist,
mathematician, and all-around scientifically and philosophically
adept individual. He asked the Dalai Lama for an opinion
on the possibility of conscious computers. The result was
the following conversation, involving Varela, two of his
colleagues, and the Dalai Lama:
DALAI
LAMA: In terms of the actual substance of which computers
are
made, are they simply metal, plastic, circuits, and so forth?
VARELA:
Yes, but this again brings up the idea of the pattern, not
the
substance but the pattern.
DALAI
LAMA: It is very difficult to say that it's not a living
being,
that it doesn't have cognition, even from the Buddhist point
of view.
We maintain that there are certain types of births in which
a
preceding continuum of consciousness is the basis. The consciousness
doesn't actually arise from the matter, but a continuum
of consciousness
might conceivably come into it.
HAYWARD:
Does Your Holiness regard it as a definite criterion that
there must be continuity with some prior consciousness?
That
whenever there is a cognition, there must have been a stream
of
cognition going back to beginningless time?
DALAI
LAMA: There is no possibility for a new cognition, which
has no
relationship to a previous continuum, to arise at all. I
can't totally
rule out the possibility that, if all the external conditions
and the
karmic action were there, a stream of consciousness might
actually enter into
a computer.
HAYWARD:
A stream of consciousness?
DALAI
LAMA: Yes, that's right. [DALAI LAMA laughs.] There is a
possibility that a scientist who is very much involved his
whole life
[with computers], then the next life . . . [he would be
reborn in a
computer], same process! [laughter] Then this machine which
is
half-human and half-machine has been reincarnated.
It is intriguing how the Dalai Lama manages to reconcile
postmodern technology with his ancient, some would say outmoded,
belief system. Like all Tibetan Buddhists, he clearly believes
that we all have a soul of some sort, which leaves the body
at death and is then reincarnated into another body. But
he is willing to consider that a computer program could
be made to manifest the abstract patterns that the voyaging
soul uses to identify a “mind,” a potential
next-nesting-place. He even conjectures that a sufficiently
advanced yogi might be able to project their soul into a
computer program: a very convenient uploading mechanism,
much less risky than the technological means we have at
our disposal today! :
ROSCH:
So if there's a great yogi who is dying and he is standing
in front of the best computer there is, could he project
his
subtle consciousness into the computer?
DALAI
LAMA: If the physical basis of the computer acquires the
potential or the ability to serve as a basis for a continuum
of
consciousness. I feel this question about computers will
be resolved
only by time. We just have to wait and see until it actually
happens.
Of
course, the Dalai Lama does not specify how, when it actually
happens, he intends to tell if a computer program truly
has acquired a mind and soul or not. The implicit assumption
seems to be that, if it talks like a mind and quacks like
a mind, then it has a mind – and if it has a mind,
it has some kind of reincarnated soul.
There
are two pesky technical details involved in the uploading
idea. One is creating a computer suitable for containing
a human mind. The other is actually getting the information
characterizing a human mind out of the human brain that
carries it. Until these problems are solved, uploading remains
a plan and a dream. There is little doubt that they’ll
be solved eventually, given the relentless advance of all
relevant technologies, but it’s not yet clear exactly
what form the solution will take.
The first of these tasks – creating an appropriate
computational substrate – is something we’ve
already talked about a lot in previous chapters. Moore’s
Law has been going on for decades and shows no signs of
abating. There are thought to be around 1010 neurons in
the brain, and 1013 -1015 synapses connecting them. According
to this, if Moore’s Law holds up, we’ll have
achieved computers with human-brain-scale memory and computing
power within just a few decades. And the wonder of exponential
growth is, even if our estimates of the brain’s memory
and processing power are low by a couple orders of magnitude,
it will only push the advent of brain-capacity computers
off by a decade or so.
The development of the process of getting the mind out of
the brain is chancier to predict, because brain scanning
is a newer technology than computing. But even so, it’s
advancing at least as rapidly. Each year the accuracy of
scanning techniques gets better and better – yielding
a Moore’s-Law-like exponential acceleration.
One promising brain-scanning technology in common use today
is MRI, Magnetic Resonance Imaging. MRI machines are a wonderful
application of fairly basic particle physics. The key phenomenon
exploited in an MRI machine is the fact that the proton
in the nucleus of a hydrogen atom, when exposed to a magnetic
field, acts like a small bar magnet and resonates. Furthermore,
the resonant frequency is proportional to the strength of
the field. So if you expose a brain to a magnetic field
whose strength varies across different regions, the protons
at different positions will resonate at different frequencies.
From this it’s possible to tell how many protons are
in a given region of space. The more blood is flowing in
a given region of the brain, the more water molecules are
there; and the more water molecules are there, the more
hydrogen atoms are there. So MRI machines track the flow
of blood through the brain, which is meaningful because
blood flow follows activity – the regions of the brain
that are active require more energy, more blood.
This is fantastic stuff. But it has a long way to go. Right
now MRI is nowhere near accurate enough to make a completely
detailed picture of a brain-state. The resolution is around
a cubic millimeter. How far the resolution can be pushed
isn’t yet clear. Some researchers have suggested radical
possible improvements, such as distributing Helium-3 through
the brain, which would diffuse rapidly into cells and permit
resolution sufficient for uploading. Also, if one is willing
to slice the brain up into little sections, it’s much
easier to use MRI with high resolution – the disadvantage
being the destruction of the brain involved.
And plenty of more radical techniques have been proposed.
For instance, if one has a stained, vitrified brain (a brain
frozen to a very low temperature using special antifreeze
chemicals that prevent ice crystal formation), then one
can abrade the brain with UV rays and use ultra-sensitive
mass spectrometry to infer brain structure. Or, alternately
but similarly, one can use an experimental method called
abrasive atomic force microscopy. These particular example
strategies are likely to be destructive of the brain as
well. But it’s not too outlandish to conjecture that
the next few decades will bring a high-resolution non-destructive
brain scan technology, be it a variant of MRI or something
altogether difference.
Digitizing
a working human brain is the holy grail of uploading research.
But science nearly always achieves its grand goals by small
steps. In this case, the first step being taken by adventurous
researchers is the uploading of a simpler organism: a nematode
worm. A nematode’s brain has only a few hundred neurons,
rather far short of the human brain’s hundred billion
or so. But it only took us half a century to get from room-sized
computers that could add, subtract, multiply and divide,
to where we are now with computing: amazingly powerful general-purpose
machines on every developed-world desktop. And the general
pace of technological advance is accelerating. Treading
the path from nematode brain to human brain may not take
all that long.
So far no one has reconstructed a nematode worm’s
brain in an entirely automatic way. Rather, human scientists
had to analyze the images of all 959 cells (338 of which
are brain cells), explicitly attending to the lineage of
these cells from the fertilized egg. But automation of this
kind of process is not too far off. For instance, researchers
have carried out a complete automated reconstruction of
a capillary bed – which is easier for technical image
processing reasons. Image processing software gets better
and better, and integration of this software with knowledge
bases embodying biological intuition of various sorts is
surely not too far off.
And although this began as a biology project, a group of
computer folks calling themselves the Mind Uploading Research
Group (MURG) has adopted the nematode upload project as
its own (http://inforezo.u-strasbg.fr/~kerner/Projects/Nematode/).
They aim to make a complete computer simulation of the dynamics
of the nematode, with an emphasis on its nervous system.
The project is explicitly seen as a first step toward creating
digital models of more and more complex organisms.
Given
that uploading a nematode worm still poses annoying technical
problems, it’s clear we’re not yet ready to
upload a human brain. But as nanotechnology pioneer Ralph
Merkle has pointed out, it’s perfectly plausible at
the present time to shoot for an only slightly smaller goal.
In his view, “At the present time, a reasonable research
objective is the fully automated analysis of a cube of complex
neuropil about 100 microns on a side.” From a nematode
worm, to a chunk of brain, and then next, to the brain as
a whole. Merkle estimates, based on a very careful and detailed
analysis, that the cost of a whole-brain-scanning research
project might be roughly in the $10 million range: considerably
less than a single advanced US fighter plane.
From a contemporary common sense point of view, the idea
of yanking a mind out of the brain it’s embodied in,
and transplanting it into a computer of some sort, sounds
completely outlandish. But yet it will require no tremendous
scientific revolutions to do such a thing. Such an achievement
is firmly within the confines of physical law, so long as
one accepts that the structure of the brain contains the
essence of the mind – a conclusion that is very firmly
indicated by all of modern neuroscience. There are many
engineering challenges to be overcome, but these are challenges
of scale – bigger computers, more accurate brain scanners.
This is not something like time travel, which is possible
only if current science is incorrect or incomplete. The
possibility of uploading is staring us right in the face,
and it seems highly probable that sometime within the 21’st
century it will become a reality.
Or in other words, as a fellow named Xiaoguang Li said recently
on the SL4 futurist discussion list: "Uploading is
a no-brainer"!
|
|
|
