[FoRK] the Singularity is far
eugen at leitl.org
Fri Jul 15 02:43:18 PDT 2011
The Singularity is Far: A Neuroscientist's View
David J. Linden at 11:15 AM Thursday, Jul 14, 2011
David J. Linden is the author of a new book,The Compass of Pleasure: How Our
Brains Make Fatty Foods, Orgasm, Exercise, Marijuana, Generosity, Vodka,
Learning, and Gambling Feel So Good. He is a professor of neuroscience at The
Johns Hopkins University School of Medicine and Chief Editor of the Journal
Ray Kurzweil, the prominent inventor and futurist, can't wait to get nanobots
into his brain. In his view, these devices will be equipped with a variety of
sensors and stimulators and will communicate wirelessly with computers
outside of the body. In addition to providing unprecedented insight into
brain function at the cellular level, brain-penetrating nanobots would
provide the ultimate virtual reality experience. In an interview with GOOD
magazine, Kurzweil says:
"By the late 2020s, nanobots in our brain, that will get there
noninvasively, through the capillaries, will create full-immersion
virtual-reality environments from within the nervous system. So if you want
to go into virtual reality the nanobots shut down the signals coming from
your real senses and replace them with the signals that your brain would be
receiving if you were actually in the virtual environment. So this will
provide full-immersion virtual reality incorporating all of the senses."
Of course, there's no reason why these nanobots must be restricted in their
manipulations to the sensory portions of the brain. In Kurzweil's scenario,
brain nanobots could just as easily manipulate motor functions, cognitive
processes, memories, emotions, and basic drives. But nanobot-mediated virtual
reality, virtual emotion, and modulated cognition are only the beginning.
Kurzweil predicts that by the late 2030s, we will be able to routinely scan
an individual's brain with such molecular precision and with such a complete
understanding of the rules underlying neuronal function and plasticity that
we will be able to "upload" our mental life into a vastly powerful and
capacious future computer. As Kurzweil describes it his book The Singularity
is Near , "This process would capture a person's entire personality, memory,
skills and history."
At that point, boundaries between brain, mind, and machine would fall away.
Once our individual mental selves are instantiated in machine form,
manipulations of mental function, perception, and action just become software
modules. Want to improve your mood? Want to preserve all your experiences in
memories with perfect fidelity? Want to have the mother of all orgasms?
There's an app for that.
As much as I respect Ray Kurzweil and appreciate his willingness to make
predictions about and argue for specific future events, I take issue with his
timetables for both the introduction of brain-nanobots and the ability to
upload the contents and meaning of a brain.
neuro2.jpg Image: Harris KM, Fiala JC, Ostroff L. Structural changes at
dendritic spine synapses during long-term potentiation.. Philosophical
Transactions of the Royal Society of London, Series B 358, 745-748 (2003).
I am a neurobiologist and I have spent the past 28 years engaged in studies
of the cellular and molecular basis of memory and cognition. I am an optimist
and a technophile, but I believe that I speak for the vast majority of brain
researchers when I express serious doubts about Kurweil's timetable.
The central premise underlying his predictions is that enabling technologies
like computer processors, computer memory, microscopes, brain scanners, and
DNA sequencing machines have been on an exponential rather than a linear
trajectory in terms of their capacity, speed, resolution, and real-world
cost, and that it is reasonable to imagine that this exponential trend will
continue. Kurzweil also assumes that the human mind resides entirely in the
brain (or at least in the nervous system): There is no immortal soul,
collective energy, or other nonbiological component that encodes our
individual mental selves. At this point in his argument I'm still on board.
However, Kurzweil then argues that our understanding of biology—and of
neurobiology in particular—is also on an exponential trajectory, driven by
enabling technologies. The unstated but crucial foundation of Kurzweil's
scenario requires that at some point in the 2020s, a miracle will occur: If
we keep accumulating data about the brain at an exponential rate (its
connection maps, its activity patterns, etc.), then the long-standing
mysteries of development, consciousness, perception, decision, and action
will necessarily be revealed. Our understanding of brain function and our
ability to measure the relevant parameters of individual brains (aided by
technologies like brain nanobots) will consequently increase in an
exponential manner to allow for brain-uploading to computers in the year
That's where I get off the bus.
I contend that our understanding of biological processes remains on a
stubbornly linear trajectory. In my view the central problem here is that
Kurzweil is conflating biological data collection with biological insight.
A Lake of Data, A Puddle of Knowledge
Let's take genetic sequencing as an example. Yes, we have now sequenced quite
a few human genomes and, yes, the speed and cost of doing so are improving
exponentially. The human genome sequence—and those of the rat, mouse, fly,
zebrafish and rhesus monkey—are an invaluable tool for biologists. That said,
while the fundamental insights that have emerged to date from the human
genome sequence have been important, they have been far from revelatory.
For example, we have learned that gene duplication is more common than we
originally thought. It's not all that rare for regions of chromosomes to
repeat themselves. We have also learned that humans have fewer genes, but
that those genes have more complex modes of regulation and more splice-forms
than we had initially predicted.
That's all useful information, but it doesn't represent a game-changing,
exponential transformation in our understanding of genetics. When the human
genome sequence was finished, no one was able to look at it and say, "A-ha,
now I can understand what makes us uniquely human," or "A-ha, now I see how a
fertilized egg becomes a newborn during the course of gestation."
There have been a number of genuine paradigm-shifting insights in genetics in
recent years. For example, we now know that chemical modification of DNA
through a process called methylation can alter its structure and the way in
which it interacts with a set of regulatory/structural proteins called
histones, thereby silencing the expression of certain genes. This is one of
several mechanisms that controls the regulation of gene expression or
"epigenetics". Such insights have explained a whole set of puzzles and are a
major step forward in our understanding of genetics.
But these discoveries, and most of the other key conceptual breakthroughs in
this field, have come slowly, the result of stubbornly linear small science,
and not of the huge technology-driven data sets that Kurzweil describes.
This linear progress also holds true for the growth in our knowledge of brain
function. For example, we now have a map called the Allen Brain Atlas that
shows the expression pattern of almost every gene in the mouse brain,
detailed in a huge series of microscopic images. This resource, which is
available to everyone on the Internet, is a wonderful tool for brain
researchers, but it has produced few "Eureka!" moments. The temporal and
spatial resolution of our brain scanners is also improving, but these
improvements have likewise yielded fundamentally linear insights.
Kurzweil's ideas about nanobots in the brain are problematic, as well.
He says his nanobots will measure seven microns across—about half the
diameter of a typical neuronal cell body—and their job will be to maneuver
through brain tissue and deploy microsensors and stimulators to evaluate
normal brain function.
You might imagine the nanobot as a car, something the size of a Volkswagen
Beetle. It drives down the road, until it finds something the size of an SUV
(a neuron). Here is the first of many problems in Kurzweil's scenario: The
brain is composed of neurons and glial cells—non-neuronal cells that
outnumber neurons 10-to-1 and provide metabolic support and slow forms of
information processing in the brain. These cells are packed together very
tightly, leaving only miniscule gaps between them.
It is easy to look at the left panel of the figure that shows a
computer-based reconstruction of the tip of a growing axon in the brain and
imagine that there is plenty of space around it. However, the complete view
of this same growing axon tip is shown in the panel on the right. This image
is made with a transmission electron microscope and it shows how the same
growing axon (marked with asterisks) is packed into a dense and complex
matrix of tissue containing other neurons and glial cells. The scale bar in
the left panel is 0.5 microns long, about 1/160th of the diameter of a human
hair. So you can imagine Kurzweil's brain nanobot, a structure about fourteen
times larger in diameter than the scale bar, crashing through this delicate
web of living, electrically-active connections.
What's more, the tiny spaces between these cells are filled not just with
salt solution, but with structural cables built of proteins and sugars, which
have the important function of conveying signals to and from neighboring
cells. So let's imagine our nanobot-Volkswagen approaching the brain, where
it encounters a parking lot of GMC Yukon SUVs stretching as far as the eye
can see. The vehicles are all parked in a grid, with only one half-inch
between them, and that half-inch is filled with crucial cables hooked to
their mechanical systems. (To be accurate, we should picture the lot to be a
three-dimensional matrix, a parking lot of SUVs soaring stories into the sky
and stretching as far as the eye can see, but you get the idea).
Even if our intrepid nanobot were jet-powered and equipped with a powerful
cutting laser, how would it move through the brain and not leave a trail of
destruction in its wake?
The nanobot also needs its own power source. And it needs to evade reactive
microglia, specialized brain cells that attack and engulf foreign bodies. And
all of this has to happen in a way that does not compromise the physiology
that the nanobot is trying to measure. These problems are not fundamentally
or philosophically unsolvable, but they are enormous. The 2020s are coming up
fast, and so there's a lot that would need to be accomplished in a very short
time to keep Kurzweil's nanobot timetable on track.
Don't get me wrong. I do believe that the fundamental and long-standing
mysteries of the brain will ultimately be solved. I don't hold with those
pessimists who claim that we can never understand our minds by using our
brains. I also share Kurzweil's belief that technological advancement will be
central to unlocking the enduring mysteries of brain function. But while I
see an exponential trajectory in the amount of neurobiological data collected
to date, the ploddingly linear increase in our understanding of neural
function means that an idea like mind-uploading to machines being usefully
deployed by the 2020s or even the 2030s seems overly optimistic.
Featured, Science • Tags: neuroscience, singularity
More information about the FoRK