All it takes is a quantum computer, an atomic-scale body map,
and a human-vaporizing laser.
By Charles Platt
Online, datastreams seem weightless and instantaneous. Down
here in the physical world, we still need constant assistance
from cars, airplanes, elevators, and our own weary muscles to
combat the pull of gravity and the drag of friction. Imagine
our lives if we could negate our bodies and become free as
bits. Imagine mass transformed into data, flowing as
electrons to a receiver that rebuilds us. Imagine
teleportation as streaming media.
The idea has a storied past. In folk tales, visions of the
disembodied were often the stuff of horror or religious
euphoria, but starting in the 1950s, matter transmission came
to be seen as a benign out-of-body experience as superscience
displaced superstition.
"American science fiction of this period depicted the
triumph of American inventor-geniuses over a world unknown
but open to conquest," says John Clute, whose
Encyclopedia of Science Fiction is the field's most
authoritative reference work. At the same time, teleportation
shifted from a supernatural function to a hardware
application.
In his haunting 1960 novel Rogue Moon, Algis Budrys proposed
a machine that could record the positions of atoms in a human
being and transmit this data to the moon, where the traveler
would be re-created from local materials.
The idea also materialized in Hollywood. The Fly dramatized
the headache of an incomplete reconstruction - messy
teleportation, you might say - while on Star Trek, the
transporter went from little more than a gimmick to an
essential plot device. Even people who didn't watch the show
know what "Beam me up, Scotty" means.
While Clute dismisses the catchphrase as "a kind of
joke, something you can say at a point where, comically, you
hope something will get you out of this," the power of
the underlying concept should not be underrated. It's a
fundamental form of wish fulfillment, one step away from
resurrection. And now that science fiction has taken some of
the horror out of transcending physical limits, if
teleporting were here today, what we'd most want to know is,
"When can I go?" and "How much will it
cost?"
At Caltech, in Pasadena, the basement of the Norman Bridge
Laboratory of Physics is like a Gothic crypt. The concrete
walls and floor appear to be of stone; the ceiling is low and
vaulted. You might expect to find a monastic cell behind this
pair of wood-paneled doors - but they part to reveal a small,
windowless laboratory crammed with exotic hardware. Here in
this subterranean den, quantum physicist H. Jeff Kimble and
his associates intend to be the first in the world to
teleport a single atom.
A 1-ton steel-topped table, specially isolated from the
tiniest vibration, is strewn with a complex clutter of
lenses, beam splitters, lasers, and cables. Above it is
clamped a double-compartment vacuum chamber that looks as if
it has been carved out of solid steel. About 1 million cesium
atoms are fed into the baseball-size upper compartment, its
interior fitted with six laser beams. "The photons from
the lasers go whack, whack, whack, removing the atoms'
kinetic energy in about one one-thousandth of a second,"
Kimble explains. Since the energy of a substance's atoms
determines its temperature, the laser light cools the cesium
atoms to an ultrafrigid minus 460 degrees Fahrenheit simply
by immobilizing them.
Next, the stunned atoms fall into the lower compartment, a
beautifully machined metal drum about 8 inches in diameter,
wrapped in coils of heavy copper wire. A porthole reveals two
glass rods that taper to delicately curved, highly polished
tips 1 millimeter in diameter and 50 microns apart. More
laser light reverberates between the rods as the atoms drop
between them. "When an atom goes through there,"
says Kimble, "it feels the impact of just one
photon."
Kimble has the dry Texas accent of a farm boy, but he talks
the language of quantum mechanics at the tempo of a Lone Star
State auctioneer. His torrent of technicality is almost
drowned out by the background roar of cooling fans, but his
ultimate goal is clear enough: to capture an atom's quantum
state and, ultimately, transmit it to another location, where
that state will be re-created in another atom. When this
happens, the atom will seem to have jumped from one place to
another, even though an observer might object that the second
atom is different from the first.
"But it's not different," Kimble says with an
enigmatic smile. "You could make any measurement, with
infinite patience, and never tell the difference."
Thus, although the quantum state is destroyed by the means of
measuring it, the state has been effectively re-created in a
second atom with such precision that, for all practical
purposes, it is the original.
At this point in the discussion, when Kimble's basement
operation starts to seem like the crude beginnings of a
cosmic photocopy machine, visitors usually ask him about Star
Trek. Patiently, he explains the huge difference between
single atoms and whole objects. Even a tiny bacterium
contains about a trillion molecules, each made of one or more
atoms. The complexity of their interrelated quantum
information is so great, Kimble insists, that the task of
teleporting a bacterium is "ridiculous on the scale of
any imaginable civilization." Doing it to a human being
would be totally inconceivable.
Still, his experiment validates the core concept of quantum
teleportation, and we may yet find ways to circumvent the
seeming impossibilities.
One man willing to entertain the impossible is Seth Lloyd. A
quantum mechanical engineer and associate engineering
professor at MIT, Lloyd calculates that if we could represent
bits of data using quantum states of electrons, all the data
in all the world's computers could be stored in a grain of
salt. He also imagines a quantum internet using the kind of
teleportation Kimble hopes to pioneer. Even Lloyd, however,
scoffs at the idea of teleporting all the quantum data
contained in a human.
"How will the initial measurement be done?" he
asks. "Theory tells us that it has to be destructive. I
have a master's in math and a master's in philosophy and a
PhD in physics, and they all tell me that whatever technique
we use, it's not going to be a pretty sight. You might become
a pile of elementary particles or a small piece of wax. It
would have no resemblance to you whatsoever."
But perhaps this messy disassembly would be acceptable if you
could be sure of being reassembled at the other end?
"All right," Lloyd says, "suppose we subject
you to an extremely intense pulse of laser light to
completely vaporize you." (This is similar to Kimble's
technique for individual cesium atoms.) "As long as we
do this carefully" - Lloyd laughs - "the vaporized
version of you, plus the light beam, still contains all the
information about you before you were vaporized. Now we take
the quantum states of the vapor. Photon by photon, we
teleport that information to the other side. Using a quantum
computer, in theory we can unvaporize you by using another
very powerful pulse of light."
And this is plausible? "No," says Lloyd. "We
have no idea in practice how to carry out those
operations." He pauses for a moment, considering the
problem. "Actually," he goes on, "I can
certainly imagine the process for vaporizing you. But the
unvaporizing process strikes me as potentially rather
difficult. The laws of physics do allow it - but they don't
encourage it."
He is willing to consider an alternative, though, if we're
willing to make some compromises.
"I can certainly imagine vaporizing you," says MIT
quantum engineer seth Lloyd. "But the
unvaporizing process strikes me as rather difficult."
Back in 1989, Donald Eigler and Erhard Schweizer pulled off
an astonishing trick, nudging individual xenon atoms on a
nickel surface to form the name of their employer: IBM. The
atoms appeared as blurry bumps, not the neat little spheres
of red and blue you see in chemistry textbooks. To nail down
their precise locations, quantum data would have been
necessary.
And yet, their positions were clear enough to form three
recognizable letters of the alphabet. Maybe there's a moral
here. Maybe the same kind of rough-and-ready information
would be good enough to reconstitute a person. After all,
there's no evidence that changing precise quantum states
damages the brain. Magnetic resonance imaging scrambles
quantum states, yet patients emerge from MRIs seemingly
unchanged.
This distinction is vital. Capturing all the interrelated
quantum states in a whole person is such an immense task that
at the very least it would require millions of years. Getting
a rough estimate of where the atoms are is a huge challenge,
but it's not necessarily impossible - and it may be close
enough for the purposes of teleportation.
Lloyd agrees that an approximate atomic scan "might be
able to re-create you with sufficient accuracy that you'd be
happy enough to pay the teleportation company."
"Maybe all we need to do is send the Lego plans for
rebuilding the molecules," says Jeff Kimble.
"That's going to take a little bit of FedEx space, but
one could imagine we could do that."
One scientist - Ralph Merkle, formerly of Xerox PARC and now
at the Texas-based company Zyvex - has taken this concept
seriously enough to do a thorough feasibility study. In a
massive 33,000-word paper, "The Molecular Repair of the
Brain," Merkle details how it could be done.
He begins by assuming that we will have molecular
nanotechnology. First proposed by the late, great physicist
Richard Feynman, nanotech involves manipulating individual
atoms. K. Eric Drexler's 1986 book Engines of Creation filled
in the blanks. John Armstrong, while he was Big Blue's chief
scientist, predicted that "we will have the ability to
make electronic and mechanical devices atom by atom when that
is appropriate to the job at hand." Drexler proposed a
molecular assembler as the basic tool of nanotech: a robot
the size of a virus, able to plug atoms together. Initially,
the assembler will make copies of itself, and then the copies
will make copies, until we have billions of them. After that,
they can do some useful work.
Unfortunately, the nanomachines will still be outnumbered by
the vast quantity of molecules in the brain. Merkle
calculates that the average brain has a volume of about 1,350
cubic centimeters and consists of about 80 percent water, 10
percent lipids, 8 percent protein, and 2 percent other
materials. From this he figures that the brain contains 1.2
sextillion protein molecules, 200 times as many lipid
molecules, and 40,000 times as many water molecules. Locating
each molecule and recording its location is going to some
time. Even with 3.2 quadrillion molecular-repair devices
working in parallel (weighing a total of about 1 pound),
Merkle estimates that to strip and rebuild a brain will
require three years.
It gets worse. The brain would have to be deep-frozen to
prevent deterioration and minimize thermal noise. The
freezing process itself would cause damage. And total energy
consumption of the nanomachines over three years would be
about 10 megawatts. Add it all up and the dream of
decapitated teleportation begins to seem more like a long,
nightmarish Victorian ocean voyage.
But Merkle has a plan B.
Some atoms are so unimportant, we don't need to know where
they are. "Describing the exact position and orientation
of a hemoglobin molecule within a red blood cell is
completely unnecessary," Merkle writes. "Each
hemoglobin molecule bounces around within the red blood cell
in a random fashion." Also, many brain components have a
simple, standardized function. Mitochondria, for instance,
feed energy to cells. No two mitochondria are precisely the
same, but, continues Merkle, "these differences don't
matter much and could reasonably be neglected." So
instead of telling our nanomachines to replicate each
mitochondrion one atom at a time, we tell them to hoist
premanufactured mitochondria into position, like two-by-fours
in a house under construction.
The simplification process can be taken further. The bottom
line, according to Merkle, is our memories, which obviously
are essential to identity. Studies by Thomas K. Landauer at
Bell Communications Research during the 1980s showed that
long-term memory storage in human beings occurs at the low
constant rate of about 2 bits per second. Therefore, Merkle
calculates, the absolute minimum to store a lifetime of human
memories would be slightly more than 100 megabytes - trivial
by modern standards. Of course, we don't know yet how
memories are encoded, and the method seems to be very subtle.
Still, copying some crucial aspects of a human brain begins
to look more manageable. As for the body - again, you can
insist on a precise replica or you can settle for something
generic. "Assuming you don't require total accuracy -
just the right number of fingers and toes - you don't need a
cell-for-cell, molecule-by-molecule replica of your body to
be quite happy with it," Merkle asserts.
An alternate option would be to upload the basic structure
and chemical states of a brain into a very large computer,
where your intelligence would live in perpetual cyberspace. I
explored this scenario in my novel The Silicon Man, which
suggested that if advanced processing power could provide
total realism in all five senses, many people might prefer to
shed their biological life-support systems and become
information entities, or infomorphs, immune to the aging
process.
The big question is, How fast must a computer be to mimic a
human brain in every detail? Using three methods of
measurement, Merkle concludes that the human brain has an
equivalent computational power of between 1013 and 1016
operations per second. In other words, we require at least
10,000 gigaflops to achieve an emulation. By comparison, the
Apple G4 promises 1 gigaflop. This sounds disappointing, but
we could run 10,000 G4s in parallel and they wouldn't cost
more than an old-style supercomputer - if Apple allows us a
bulk-purchase discount.
Alternatively, if computer power continues to double every 18
months, we should have a desktop computer with the power of a
human brain about 20 years from now. After uploading yourself
into it, you could transmit electronic copies of your brain
via any conventional link: radio waves, fiber optics, or
twisted-pair telephone wire - although that might take a
while. Clearly, a downloadable brain would be a real
bandwidth hog, and error-correction protocol will be
mandatory if you don't want to end up like the unfortunate
doctor in The Fly.
But none of these far-fetched speculations violates any
physical laws. Merkle's strip-and-rebuild scenario, in
particular, has the potential for realization within the next
half-century. In the words of IBM researcher Donald Eigler,
"By the time I'm ready to kick the bucket, we might be
able to store enough information on my exact physical makeup
that someday we'll be able to reassemble me atom by
atom."
When will the basic tools of nanotech be ready for the job?
"If we did a massive parallel effort with an infinite
supply of money, we might be able to build a basic assembler
in 10 years," Merkle says. "Realistically, 20 years
looks like a more plausible guess. But certainly it will be
decades rather than centuries."
After uploading yourself, you could transmit electronic
copies of your brain via any conventional link: radio, fiber,
even twisted pair.
Suppose we have the power one day to rebuild a person from a
new set of atoms. Will it be the same person?
Merkle dismisses this question, pointing out that constant
atomic replacement occurs naturally in every human being.
"It does not appear that substituting atoms changes who
you are," he says. "If you eat a Big Mac, you
continue to be the same person, even though normal
biological-repair processes use the molecules from that
dinner to replace pieces of tissue."
Nonetheless, there will be profound questions about identity,
since our sense of self is a purely internal phenomenon that
cannot be measured or verified. Captain Kirk and his crew had
occasional misadventures along these lines, and Algis Budrys
explored the question at length. In Rogue Moon, the
protagonist asked himself whether the cover of his
schoolbook, decades before, had been red or blue. Suppose he
feels certain that it was blue. Then his mind is
destructively scanned and re-created, and now he is sure that
the book was red. This inaccuracy could be introduced by the
transmission - but the new person is just as confident of his
memories as the old iteration was, and there's no way to
settle the matter, since the brain cannot verify whether a
memory is true or false.
We should expect no guarantees. Teleported people who claim
to feel the same as before may be fooling themselves. Others
who complain that they feel different may be suffering from
hypochondria. None of them will know for sure.
There are obvious questions, too, about the human spirit. In
fact, successful matter transmission could render obsolete
the concept of a soul.
"As a scientist, I am not skilled in the area of
discussing the soul," Merkle says, "but I can point
out that tens of thousands of people alive today were frozen
as embryos, and they appear to be none the worse for
wear." In other words, life processes can be interrupted
- and can resume - without apparent damage to the spirit. If
the soul exists, it might survive a brief transition through
the limbo of matter transmission just as victims of
cold-water drowning can revive hours after the loss of vital
signs.
On the other hand, if we can rework a lump of carbon into a
human copy indistinguishable from the original, this might
trivialize life itself, threatening our assumption that each
of us is a unique, irreplaceable individual. Anyone peddling
such a procedure will be a target for backlash, much as
Galileo was punished for challenging the ancient conceit that
Earth sits at the center of the universe.
Even more problems arise if we use the atomic data defining a
person to make multiple copies of that person. Playfully,
Seth Lloyd imagines the consequences if matter transmission
is not properly secure. "Nothing would prevent someone
from intercepting the transmission, stealing the atomic
information, and making another copy of you," he says.
"It could go around running up huge credit card debts,
making love to your wife, and doing other things you might
not want it to do, while you are on some other planet doing
serious research." Clearly, building consumer confidence
in teleportation will take some doing.
Fortunately for entrepreneurs, many of the issues facing a
teleportation startup can be anticipated: Since a molecular
scan of a large object almost certainly will entail
disassembling the object before it can be reconstructed,
perform beta testing first on simple inanimate forms. If that
works, test it on lab animals. (Prepare a defense for the
predictable protest from People for the Ethical Treatment of
Animals.) Finally, offer the service initially to humans only
as an alternative to death by terminal disease.
Next, establish a market for molecular repair by treating
accident victims whose bodies have been irrevocably damaged
but whose brains are intact. A few successful body
replacements should win many more clients, especially if
artificial intelligence controls the process. A human clerk
could never copy millions of numbers without making
transcription errors, yet today we routinely trust computers
to do this for us. In the future we will place a similar
degree of trust in systems performing molecular
transcription. Looking ahead, as the perceived risk
diminishes, the benefits of human matter transmission will at
some point seem to outweigh its dangers, and the procedure
will gain popular acceptance, just as air travel - once
perceived as a horribly risky adventure - is now a dull
routine endured by millions of business travelers.
Crime could become uncontrollable, but low-cost matter
transmission would render material wealth, objects, and
far-off places instantly and ubiquitously accessible;
national borders, currencies, and cultures could well become
irrelevant.
On a positive note, matter transmission would free us from
our planet's confines by enabling low-cost space travel as
envisioned by Algis Budrys in Rogue Moon. First, a
conventional unmanned rocket would soft-land a teleportation
receiver and nanotech mining equipment on the moon to extract
raw materials from local rock. Next, we could scan a person
on Earth, transmit the data, and rebuild a human copy 240,000
miles away. Using the same technique, we could colonize
nearby planets and asteroids.
At present, of course, we haven't reached square one. Jeff
Kimble and his associates can capture the quantum state of a
single atom, but they haven't teleported it yet, and their
quantum-optics facilities at Caltech took 10 years to build.
"My group and I feel like the juggler with the plates
and the sticks," Kimble says. "You get one plate
going, and then you get that one going, and that one, and you
have to go work on the other one - and then a plate falls on
the floor. We're just running and running and turning those
sticks and plates."
If the 20th century has taught us one thing, it is to be
cautious about the word impossible.
Darpa has helped finance the work, together with the National
Science Foundation, the Office of Naval Research, and
Hewlett-Packard. But some prospective funders are impatient
with the decades the research requires.
As Kimble talks of mundane topics such as money and staffing
amid the delicate, expensive equipment in his laboratory,
visions of disassembling and reassembling people begin to
fade. Still, if the 20th century has taught us one thing, it
is to be cautious about the word impossible. Technology
already has freed us from limits that used to be considered
absolute. Freeing ourselves from the ever present burden of
our own physical mass may just take a little longer.
Senior writer Charles Platt ([email protected]) wrote about
wireless broadband in Wired 7.12.
(http://www.wired.com/wired/archive/8.01/teleport.html)