[Sci.nanotech] Re: Interview with John Robert Marlow on the Superswarm Option
John Marlow
To respond to Dr. Pusch:
Among the many sourced quotes contained on my website
(www.johnrobertmarlow.com) is the following: *A single nanoweapon or
nanoaccident could "reduce the biosphere to dust in a matter of days" [8f], if
not literally destroy the planet. [9f]* (Notes refer to website links, not this
post.]
The source of the "matter of days" quote is Eric Drexler--who makes the same
analogy chosen by Dr. Pusch (bacteria), but comes to a far different conclusion
("Tough, omnivorous "bacteria" could out-compete real bacteria: they could
spread like blowing pollen, replicate swiftly, and reduce the biosphere to dust
in a matter of days"--source: Engines of Creation, chapter eleven
http://www.foresight.org/EOC/EOC_Chapter_11.html).
Thus--far from "paying too much attention to the overblown and scientifically
inaccurate claims of doomsday science fiction writers (many of whom are
scientifically illiterate luddites)," as Dr. Pusch suggests--I was in fact
relying on what Eric Drexler had said. Which is, as it turns out, also what
Robert Freitas says in the very paper cited by Dr. Pusch in support of his own
contention that such an event would take days or weeks (in the case of a city)
or two years (in the case of the biosphere) to transpire.
While it is true (as stated more than once in the NANO novel) that the maximal
rate of nanoswarm expansion cannot be sustained beyond the first instant (else
the earth and, indeed, the entire universe would be consumed in short
order)--maximal expansion throughout the swarm is not required in order to
destroy a city in minutes (which is what Dr. Pusch is apparently referring to
here when he states: "the response time required to deal with a rogue nanotech
infestation will _NOT_ be on the order of "minutes" as you claim, but on the
order of days or weeks --- i.e., comparable to the response time required to
deal with an epidemic outbreak of a typical infectious disease").
In his paper Some Limits to Global Ecophagy by Biovorous Nanoreplicators, with
Public Policy Recommendations: http://www.foresight.org/NanoRev/Ecophagy.html),
Freitas holds that the limiting factor which determines the maximum dispersal
velocity of small nanites ("biovorous nanoreplicators") is wind speed-which he
cites as being (global mean wind speed) ~10 meters per second. This yields a
swarm expansion rate of 1 kilometer per 100 seconds. A quick look at a map will
reveal that San Francisco (a rather large city, and site of the climactic scene
in the NANO novel) would, at that rate, be doomed in minutes and not--as stated
by Dr. Pusch--hours or days. Furthermore, wind speeds in coastal cities can be
considerably higher than average, making the figure just cited conservative. It
should also be noted that Freitas' figure applies only to nanites which have
NOT been specifically designed to maximize airborne dispersal rates ("isolated
replibots lacking significant aeromotive
capabilities").
Next, Dr. Pusch asserts that "Dr. Robert Frietas has shown that the maximum
possible "biomass to nanobot" conversion rate of self-replicating nanobots
cannot possibly be greatly larger than that of already-existing
self-replicating organic lifeforms such as bacteria, so that far from
"destroy[ing] the planet in a matter of days," the timescale for complete
"ecophagy" (conversion of the entire biosphere into nanobots) is in fact quite
long --- on the order of _TWO YEARS_, not two days."
Actually, Freitas' paper says nothing of the kind. On the contrary, the paper
specifically states (just below Table 4) that (as CRN co-founder Chris Phoenix
pointed out in an earlier response to Dr. Pusch's post) "a moderately efficient
goo, not trying to avoid detection, could in theory convert the biosphere in
2.44 days."
This with energy dissipation at 10MJ/kg. Freitas also points out that both Eric
Drexler and Ralph Merkle have conjectured that efficiencies on the order of 0.1
MJ/kg may be possible in some circumstances, and refers to this as "a
possibility that should be investigated in the present context in a future
theoretical study." Such efficiencies would make 2.44 days seem an eternity.
Further, this is the time required to convert the entire biosphere-and as Chris
has pointed out to me, "if we reach the 2% goo stage globally, or even the
0.02% stage, we're probably already doomed."
Freitas' paper largely concerns itself with describing what is-in comparison
with what is possible-slow-motion goo which expands at a limited rate in order
to postpone swarm detection. The scenario presented in NANO, Superswarm, and
the interview cited by Dr. Pusch, on the other hand, deals with the maximum
possible rate of swarm expansion.
Or does it? For, as Chris also pointed out to me, swarming omnivorous nanites
which convert the environment to goo (and copies of themselves) may not be the
most dangerous or the most swiftly-spreading nanohazard. That dubious honor may
belong to something which simply destroys things and moves on without massive
self-replication.
In any event, the scenario employed in NANO also involves a seed AI which is
involved in the design, production, and dispersal of nanites (this is the
"wildcard" to which Chris refers in his earlier post). Needless to say, a true
AI will rapidly determine and implement maximal efficiencies which humans may
be incapable of determining, much less implementing, in any meaningful
timeframe.
I'd like to conclude by saying that Dr. Freitas' paper appeared after the
initial draft of NANO was completed (it takes a LONG time to get a book
published); nonetheless, even if the specific technical details of the rogue
swarms present in novel, superswarm paper, and interview are in some way
incorrect-it seems clear that the same (and perhaps worse) dangers are well
within (to borrow a phrase) the limits of the possible. (Chris Phoenix: "I
don't think it would change the book much if he had written it to conform to
reasonable physical limitations. Less chemistry, more micro-scale robotics...
and you'd get much the same impact.")
When it comes to novels--tech-thrillers in particular--the practice which makes
for the most exciting read and the broadest appeal is to push things to the
limits of the possible. That doesn't mean what is possible today; it does mean
engaging in some degree of reasonable extrapolation.
Scientific papers, while important, are read by few; novels are read by many.
My aim in writing NANO was and remains to create a greater public awareness of
nanotechnology and its staggering potential to benefit Mankind--and to destroy
us. I think everyone here can probably agree that increased awareness of the
pros and cons is a good thing, and it is my profound hope that this novel will
accomplish that.
Readers whose interest is sparked by the book are directed by it to my website
(www.johnrobertmarlow.com)--which links to a growing number of nanotechnology
resources (including sci.nanotech and the Center for Responsible
Nanotechnology), which readers can explore to learn more, form their own
opinions--and act on them.
That-and not which specific details of a future technology will yield the
result Freitas, Drexler, Merkle, Chris Phoenix and others all seem to agree is
possible-is, in my opinion, what matters most in the context of a work of
plausible fiction.
---John Robert Marlow
Gordon D. Pusch wrote:
Mr. Marlow's book and interviews _GROSSLY_ overstate the maximum possible
rate that rogue nanobots can replicate. Based on extremely general physical
principles, such as the conservation of energy and the limits of available
energy vs. the characteristic energies of chemical bonds, Dr. Robert Frietas
has shown that the maximum possible "biomass to nanobot" conversion rate of
self-replicating nanobots cannot possibly be greatly larger than that of
already-existing self-replicating organic lifeforms such as bacteria, so
that far from "destroy[ing] the planet in a matter of days," the timescale
for complete "ecophagy" (conversion of the entire biosphere into nanobots)
is in fact quite long --- on the order of _TWO YEARS_, not two days.
In particular, exponential growth of a nanite infestation cannot possibly be
sustained past the earliest, initial stage of the infestation, due to energy
and resource limitations, as well as production of waste heat by the nanobots;
hence, late growth can only occur at polynomial rather than exponential
rates, and at the surface of the expanding colony rather than throughout
its volume. Therefore, the response time required to deal with a rogue
nanotech infestation will _NOT_ be on the order of "minutes" as Mr. Marlow
claims, but rather on the order of days or weeks --- i.e., comparable to
the response time currently required to deal with an epidemic outbreak of a
typical infectious disease.
Please see <http://www.foresight.org/NanoRev/Ecophagy.html> for
Dr. Frietas' complete analysis.
In short, Mr. Marlow has paid too much attention to the overblown and
scientifically inaccurate claims of doomsday science fiction writers
(many of whom are scientifically illiterate luddites), and too little
attention to the physical limitations imposed on _ALL_ forms of nanotech
by real-world physical laws.
-- Gordon D. Pusch
john robert marlow
author: NANO
"Marlow's debut is a real page-turner." --Kirkus Reviews
"Plausible, scientifically accurate, and timely....
the most important piece of fiction written to date."
--Rocky Rawstern, Editor, Nanotechnology Now
excerpts and more at www.johnrobertmarlow.com
"John Robert Marlow" <john_q...@yahoo.com>
Anton Vredegoor
On 25 Feb 2004 17:28:28 GMT, John Marlow <john_q...@yahoo.com>
wrote:
>In any event, the scenario employed in NANO also involves a
>seed AI which is involved in the design, production, and dispersal
>of nanites (this is the "wildcard" to which Chris refers in his
>earlier post). Needless to say, a true AI will rapidly determine
>and implement maximal efficiencies which humans may be incapable
>of determining, much less implementing, in any meaningful
>timeframe.
An AI with nanite designing, production and dispersal capacities poses
a real challenge to a novelist. The AI could upgrade or rebuild itself
even while expanding. So a more realistic picture would be an
expanding swarm of nanites of different generations, where more
advanced nanites would be launched later. So it would look like a
pulsed expansion where each 3-d layer (something like the layers of an
onion) would represent a technological breakthrough.
Since more advanced nanites would be able to expand faster than
earlier versions, the outer layers would be overtaken -eaten- by the
newer ones. Maybe the AI develops the earlier nanites especially so
that they can be absorbed easily. The maximally possible expansion
rate would possibly be the speed of light, but because of heat
dissipation problems only sparsely distributed probes would shoot out
of it with speeds approaching the speed of light. Maybe I'm describing
some of these immensely powerful astronomical objects that can been
already?
>When it comes to novels--tech-thrillers in particular--the practice
>which makes for the most exciting read and the broadest appeal is to
>push things to the limits of the possible. That doesn't mean what
>is possible today; it does mean engaging in some degree of
>reasonable extrapolation.
Yes, but this is more or less why I have given up on reading science
fiction. With the current rate of technological progress it's almost
impossible to publish a tech novel before it is overrun by new
scientific developments, sometimes even stranger than science fiction.
>From what I gather from other people about your book you have struck a
balance between what could be possible and what people will believe,
for which you have my admiration.
This still doesn't mean your novel is bold enough to come close to
what will happen, because you didn't -as it seems to me from here
without having been able to read your book- take into account the AIs'
potential for exponential upgrading of its own hardware and software,
which seems a very likely development to me, once the combination of
AI and nanotech exists.
Anton
Robert V Hill
I hate to say this, but this super fast Nani swarms ideal is BS. At less for
the see able future, and maybe always. First we do not have true AI yet, and
no one known when we will ever get it. It is more then a hardware problem it
is also a software problem. Some of the software problems we simple do not
have a clue how to do them like massive multi- threading. We have problem
working with more then 2 threads at once. Speed and heat are problems, the
more speed you have the more energy you need the more heat that is
generated. For any object with mass to travel at the speed of light would
take infinite energy. Also the faster a object travel the greater it mass
is. The biggest problem is energy itself were does all this energy come
from? It can't come out of thin air. Are the nanites using Solar energy? If
so then they only have a limited amount to work with, it a lot, but not that
much. Nuclear powered nanites might have more avaiably power, but they may
not be plausible, radiation is a problem for things made out of atoms.
"John Marlow" <john_q...@yahoo.com> wrote in message
news:c1ilv...@enews2.newsguy.com...
>
> To respond to Dr. Pusch:
>
>
>
> Among the many sourced quotes contained on my website
> (www.johnrobertmarlow.com) is the following: *A single nanoweapon or
> nanoaccident could "reduce the biosphere to dust in a matter of days"
[8f], if
> not literally destroy the planet. [9f]* (Notes refer to website links, not
this
> post.]
[ Excess quoted material elided by moderator. -JimL ]
Eivind Kjorstad
Robert V Hill wrote:
> I hate to say this, but this super fast Nani swarms ideal is BS. At
> less for the see able future, and maybe always.
Refering to works of fiction as "bullshit for the forseeable future" is
more than a little strange. By adding the qualifier, you yourself admit
that it migth be one possible scenario, that it's not certain that
it'll ever happen (or even be possible) is a given with any fiction
dealing with the future.
> First we do not have true AI yet, and no one known when we will ever
> get it.
True. But assuming that you consider humans to be intelligent, and
assuming that you don't subscribe to some supernatural "uniqueness" to
the human mind, it's a given that artificially constructed
intelligences are possible.
> software problems we simple do not have a clue how to do them like
> massive multi- threading. We have problem working with more then 2
> threads at once.
But any threaded system can be simulated by a non-threadened one, simply
by running one thread at a time. It would run more slowly this way, but
that is not a fundamental concern, and could be overcome by sufficient
raw power. Indeed that is how this computer, rigth now, runs 73
different processes in spite of physically only being capable of
running one at a time.
> Speed and heat are problems, the more speed you have
> the more energy you need the more heat that is generated.
True. But this problem is no more fundamental for artificial systems
than it is for natural systems. Thus natural (biological if you prefer)
systems serve as a lower limit of what is possible.
> For any object with mass to travel at the speed of light would take
> infinite energy. Also the faster a object travel the greater it mass
> is.
Now you are being difficult on purpose. All the poster was suggesting
was that it's possible the only ultimate limit on the spread-rate of a
sufficiently advanced nano-swarm is the speed of ligth itself. It makes
no real difference to the time it takes (from the point of view of say,
for example, us) to colonize the galaxy if it happens at C or at 0.9*C
> The biggest problem is energy itself were does all this energy come
> from? It can't come out of thin air. Are the nanites using Solar
> energy? If so then they only have a limited amount to work with, it a
> lot, but not that much.
It is about two orders of magnitude more than the curent biological
systems on earth manage to take advantage of.
Sincerely,
Eivind Kjrstad
John Marlow
Freitas' conservative figure projects possible complete nanoswarm-initiated
biospheric destruction in under 3 days, with no AI involved. Both Drexler and
Merkle suggest that nanite efficiencies far greater than those used in Freitas'
calculations are possible; Freitas himself recommends further study in this
area. Chris Phoenix has pointed out that we'd probably all be doomed at the 2%
(or even 0.2%) biospheric destruction level--and also that there are likely
faster ways to destroy the biosphere with nanites of a non-gray goo variety.
Personally, and in consideration of the possible consequences--I feel it unwise
to discard their opinions.
A nanite swarm could modify itself on the fly to take best advantage of
available energy sources.
The trouble with the forseeable future is that it usually isn't.
john robert marlow
www.johnrobertmarlow.com
Robert V Hill <t.bla...@comcast.net> wrote:
I hate to say this, but this super fast Nani swarms ideal is BS. At less for
the see able future, and maybe always. First we do not have true AI yet, and
no one known when we will ever get it. It is more then a hardware problem it
is also a software problem. Some of the software problems we simple do not
have a clue how to do them like massive multi- threading. We have problem
working with more then 2 threads at once. Speed and heat are problems, the
more speed you have the more energy you need the more heat that is
generated. For any object with mass to travel at the speed of light would
take infinite energy. Also the faster a object travel the greater it mass
is. The biggest problem is energy itself were does all this energy come
from? It can't come out of thin air. Are the nanites using Solar energy? If
so then they only have a limited amount to work with, it a lot, but not that
much. Nuclear powered nanites might have more avaiably power, but they may
not be plausible, radiation is a problem for things made out of atoms.
"John Marlow" wrote in message
news:c1ilv...@enews2.newsguy.com...
>
> To respond to Dr. Pusch:
>
>
>
> Among the many sourced quotes contained on my website
> (www.johnrobertmarlow.com) is the following: *A single nanoweapon or
> nanoaccident could "reduce the biosphere to dust in a matter of days"
[8f], if
> not literally destroy the planet. [9f]* (Notes refer to website links, not
this
> post.]
[ Excess quoted material elided by moderator. -JimL ]
_______________________________________________
sci.nanotech mailing list
sci.na...@nano-tek.org
http://venusia.golgothe.net/mailman/listinfo/sci.nanotech
john robert marlow
author: NANO
"Marlow's debut is a real page-turner." --Kirkus Reviews
"Plausible, scientifically accurate, and timely....
the most important piece of fiction written to date."
--Rocky Rawstern, Editor, Nanotechnology Now
excerpts and more at www.johnrobertmarlow.com
"John Robert Marlow" <john_q...@yahoo.com>
..
---------------------------------
Do you Yahoo!?
Get better spam protection with Yahoo! Mail