From: Eric Watt Forste (arkuat@idiom.com)
Date: Wed Nov 17 1999 - 15:39:17 MST
I wrote:
> > Yes, but if you work away from the "light of day" you slow down
> > the process of diffusion of innovation. Many labs will be
> > reinventing wheels unnecessarily.
Robert J. Bradbury writes:
> Not if you have a top, down distributed effort. Private development
> is usually aware of what public development is occuring but the
> converse may not be true. Private development would borrow and
> steal from everything -- government & university labs, NASA, patent
> filings (after all patent law is difficult to enforce in space and
> why would you care anyway if you are in emergency "escape" mode?).
On what basis are you assuming that all worthwhile private researchers
will join *your* research effort? I'm assuming private, nonpublishing
research efforts will be fragmented, and without regular publication and
discussion in open forums, many labs will be entirely unaware of the
existence of many if not most non-publishing labs.
> You do have to have enough secure communication between the groups
> to provide redundancy (in case one approach fails) without too
> much duplication.
Communication is only as secure as its participants. Such a network will
be extraordinarily leaky, since you will doubtless be employing many
scientists trained in the conventional, publication-oriented style of
scientific research which currently dominates all fields of research.
> > I don't know why you want to slow down the process with secrecy.
> Only to the degree that there are overhead costs associated with
> secrecy does it slow things down. You might lose valuable
> insights or criticisms into fruitless approaches if you have
> wide public review, but things are moving so fast now that
> few scientists bother to publish negative results or waste
> their time telling others that something might not work.
> > There is a *reason* why publication is recognized as a fundamental
> > part of the process of scientific research.
> Publication works in *scientific research* because you want verification
> that the conclusions are correct. There is a *big* difference between
> engineering (which is most of what Greg wanted) and scientific research
> (I've personally learned this the hard way). In engineering all you
> need to know is whether or not it works. Reality provides a much
> better review than any collection of your "peers" (rockets that don't
> work blow up). The only point on which we may disagree is the amount
> of "engineering" vs. "research" that might be required to fullfill
> the goals of getting a sustainable enclave into space.
> >
> > If you slow research down, people will die unnecessarily as a
> > result. I can't countenance that, myself.
> Well, then, I suspect you are up in arms about the drug industry
> and the FDA, since their current approaches do kill many (due to
> the unavailability of drugs that could work for a subset of the
> population) in the name of protecting a few (for whom those drugs
> may be dangerous).
Of course I'm up in arms about that, and have been for decades. But this
is utterly irrelevant.
Apparently you think that *all* the necessary basic scientific
research required to establish a mind-producing society offplanet
has already been done, and that only engineering work is needed
now. I wish that were the case, but given that none of this stuff
has been tested yet... I suspect that taking a technology from the
theoretical applied science phase into the engineering phase will
generally be the best way of turning up flaws in existing theory,
and that at least a few fundamental scientific results (with
according modifications to scientific theory) will turn up in the
course of the engineering work.
Frankly, I doubt your contention that secrecy will not slow down
the engineering work (although you seem to concede that it would
slow down work in basic science), but our difference does seem to
come down to my expectation that implementation of what are now
only sophisticated designs on paper *will* lead us to discover new
things about the structure of matter, just as it has throughout
our history as a science-practicing species, whereas you seem to
expect (for what reasons I know not) that moving a whole fundamentally
new technology from the drawing board into reality will result in
*no* important new discoveries about the structure of matter.
> I think that the process of peer review for paper publication may
> slow things down just as much as any secrecy efforts might. What
> you want is "trusted" reviewers that can give you feedback without
> the overhead & delays the publication process typically requires.
I was arguing against secrecy, not for peer review. People can
publish their own stuff on the web, and peers can decide whether
to link to it or mirror it or not. There is no need for peer review
to slow down publication... it's only a credentialing mechanism,
after all.
> > Botulin toxin can be produced easily by biotech, dissolved into a
> > DMSO vapor, and dispersed by a variety of means. Etc, etc, etc.
> > Rather than to Gray Goo, I might refer to Nick Szabo's essay on
> > Green Goo.
> Ah, but if this were perceived as a "real" threat (say the
> bioweapons treaties were never signed by a majority of countries)
> do we not have the technology to develop antisera or perhaps even
> vaccinate against or develop drugs that interfere with B-toxin?
> All of the fear of B-toxin was developed in an era when we knew
> nothing of molecular biology. In those days suppression of the
> bacteria was all you could do. Now-a-days it might make more
> sense to vaccinate everyone against it or simply add engineered
> bacteria that secrete a B-toxin binding & sequestration protein.
Uh, vaccination is a way of activating the immune system to respond to a
living organism. Botulin toxin is a molecule, not an organism. Perhaps
you will find a way to anull the toxicity of cyanide too? That would be
a neat trick. B-toxin kills more quickly than bacteria could
sequestrate it. Yes, yes, eventually you might have an active defense
in your bloodstream, ready and waiting, but post facto administrations
of nanotech would not be timely enough to help here.
But all of this is beside the point, since people will soon be
designing new toxins as fast as defenses against the traditional
ones are developed. I picked the B-toxin in DMSO example because
it's a weapon that could be developed for trivial time and cost in
any prokaryote biotech lab. (Note, prokaryote biotech is *much*
cheaper and simpler than the eukaryote biotech we see in the
prolongevity news.)
> I think these things are threats because we haven't focused any
> attention for developing defenses against them, not because we
> cannot do so.
Oh, I agree. But that doesn't mean that these things are not currently
threats, and it doesn't mean that they will not remain threats until
sometime after defenses have been developed AND tested.
> If you have a pointer for the Green Goo essay I'd like to review it.
http://www.best.com/~szabo/green.goo.html
I highly recommend the other papers at Nick's site also; his work on
smart contracts is far more important than this essay, but this essay
just happened to be relevant to the discussion at hand.
> > I do not see any mention by you of means by which you will
> > prevent the emergence of the two situations that you admit would
> > give us cause for concern. Amoral intelligences already exist
> > in plenty, and will start enhancing themselves in short order if
> > they have not already.
> Yes, but they are not operating completely unseen by us or without
> our anticipating their possible strategies and developing appropriate
> countermeasures. The real dangers lie in things we cannot see or
> do not suspect.
I beg your pardon? How are you going to get access to top-secret
military research conducted by the US Federal government?
> > And there are people whose posthumanism
> > extends as far as antihumanism... such people may already be
> > engaged in fast self-evolution research.
> While I may be strongly posthumanistic, I doubt I'm antihumanistic.
> I do not envision a situation in which as a post-human SI, I would
> really care to eliminate the primitive inhabitants of Earth (though
> I might give strong consideration to converting it into a more
> efficient simulation form). :-)
Yeah, *you* wouldn't. Neither would I. So what? I wasn't talking about
either one of us.
> My suspicion would be that competitive pressures & awareness of
> the feasibility of conscious evolution would drive some subset
> of the human species in the SI direction and the rest would
> eventually die out. Why be actively anti-humanistic, when I can
> simply be passive and let nature take its course?
Okay, but you still aren't convincing Mr. Bad over there, who grew
up in a Eurasian battlefield and has since seized a small national
government and used it to obtain a few large multinational
corporations. He's been kicked around since birth, enjoys human
company only to tyrannize it, and generally thinks that human beings
stink, with (in his case) pretty good reason.
> > I wrote:
> > > With regard to the stale atmosphere issues, these are taken care of by
> > > molecular sorting. With regard to food production, people seem to be
> > > still thinking ancient-tech. A full complement of nutritious & tasty
> > > gruels with all the necessary vitamins, proteins, fats & carbohydates
> > > will be able to be produced in bacteria/yeast, directly from sunlight
> > > or electricity, using biotech, long before we have hard nanotech
> > > available.
> >
> > I see zero economic pressure for the development of such technology.
> > Who will fund the research and why would they bother to fund it?
> Its already being done. There are 60+ complete genomes currently
> known and the funding to unravel the functions of the unknown
> genes in those genomes exists (gotta give the Ph.D. candidates
> something to do ...). It will not happen overnight, but will
> move forward quickly as more automated methods become available
> and we become more clever about the experimental and computer
> methods that can be applied to these problems.
Uh. Genome sequencing is hardly the same thing as producing and
testing wholly artificial diets for indefinite use. And the only
test of any of this stuff that will do for an engineering trial
before taking the system off into orbit to live in it indefinitely
is closure, long-term carefully monitored ecosystem closure. Again,
I see zero economic pressure for the development and testing of
this technology. Genome sequencing isn't even remotely related to
what I was talking about. You were talking about molecular sorting...
how do you get from there to genome sequencing?
> The economic pressure is clear, it is a population of ~10 billion
> people in the not to distant future (that argument was presented
> to me several days ago by the Dep. Director of the Inst. for Agric.
> Biotechnology in Moscow). The problem is that the pressure doesn't
> exist much in the developed countries where we can feed ourselves.
> In the lesser developed countries there are fewer resources and less
> development in the labs, but what they lack currently will be balanced
> in the future by shear #'s of inexpensive hands and population/economic
> throw weight (as well as the fact that they can rapidly leverage the
> technologies & information we have so painstakingly assembled).
Well, that's fine, but how come I don't see any market going on right
now? Where are the closures? You're promising me the economic pressure
will develop, but I've heard similar arguments in the early 1970s that
as soon as the planet's population hit six billion, there would be a
frantic push for space. I'm unconvinced. I've read Julian Simon; frankly
I think population "pressure" will just make things even easier for us,
so we can procrastinate space development until long after the means of
inexpensive destruction of the current (only) biosphere have been
developed and deployed.
If you want to have this technology available (by which I mean
thoroughly tested in real physical situations, *not* just in computer
simulation) in ten or twenty years, when we will probably need it, we
need to start developing it *now*. Yet, right *now*, there is zero
economic pressure for its development and testing. This is a problem.
> > I'm still not convinced that effective in silico simulations of
> > ontogenesis can be done without atomic-level detail, and that is
> > a pretty tall order.
> Agreed, you gotta get into space and build one darn big supercomputer.
Uh, yeah, but we wanted to do the simulations so we could get into
space, didn't we? Cart before the horse and all.
-- arkuat
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:05:47 MST