From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Sat Jan 25 1997 - 04:49:04 MST
On Fri, 24 Jan 1997, Omega wrote:
> Eugene Leitl wrote:
>
> > [ killjoy, much yet too little happening ]
> > Maybe we need better PR.
>
> I agree only to the extent that the things you mention reduce to human
> behavior, which is what I consider to be, by far, the biggest factor in
> this whole process, but not with what you characterize as the hardware
> requirements.
Currently, only humans are doing research. Machines are but tools
(growing smarter, granted). Cancel the humans from the equation now, and
everything will come to a crashing halt. What we need is the
sustainability of R & D progress (wonderful term, this) in the absence
of of humans. In a pinch, a pretty smart von Neumann ecosystem should
suffice, after we're gone.
Hardware requirements for a human equivalence _are_ exorbitant. Do your
back of the envelope, semiconductors won't do. Molecular circuitry is
required, which is not easy to do at all.
> > [ barriers ]
>
> I agree that the future may well be uncertain (although the possibility
> exists that it may actually be fully deterministic in light of the new
That has zero significance if I can't predict it. A
theoretical/metaphysical result merely.
> transactional interpretation of QM [a subject for the free-will thread])
> but I don't agree with your assessment of the relevant technology.
>
> Wormholes? Relativistic physics? Whoa, I know I mentioned the pos-
While wormholes is lunatic fringe technology, our current circuitry is
already constrained by (relativistic) physics. Do your back of the envelope,
it's really revealing.
> sibility of hyper-exponential development, but my original context was
> sex changes within the context of nanotech designer life. I agree there
I presume we are talking about uploads, which has excellent chances to be
the only life in existance. Having physical bodies appears to me as an
anachronism, a waste of resource. You are sure you are able to protect
your physical system if the ground you tread on will be disassembled to
become a part of the Dysonian orbiting computer cluster?
> may be fundamental barriers, but I see these as being much further out
> than anything involved in simply getting to nanotech and designer life.
Drexlerian molecular manufacturing by mechanosynthetic means can be
classified as speculative engineering. Not lunatic fringe, no, yet firmly
speculative.
> I see no reason at all why early trans-human intelligences should press
> the limits of physics anymore than the human brain itself does.
The human is an instance of a naturally evolved nanosystem. Do not sneer
at these, since such systems use molecules and cluster of molecules for
building blocks. You can improve them, but not by awful lot. E.g. a human
processing eqivalent may occupy a volume range from a sugar cube to an
orange, while requiring drastically more power to run than its biological
counterpart. While speedup factors of up to 1k might be realistic (I
haven't estimated this too exactly yet), factors of 1 M and more are
definitely not.
> It's not at all clear to me that developing nanotech even requires
> trans-human intelligence, and even if it does, I'm not at all con-
It requires a lot of R&D G$$'s, several decades, a number of
breakthroughs. What is transhuman? In 20-30 yrs, resources commanded by a
human, genetically identical to us will be quite formidable. On the
whole, a number of such individuals will command mindboggling powers. It
will be fun.
The trouble with (Drexlerian) nanotech: we don't know whether it is
feasible. Bootstrapping will be terribly hard, like stirring a cup of
cofee with a tree stump. Other, weaker brands of nanotech are possible,
but these are not very suitable for all-purpose molecular manufacturing.
(Sufficient for molecular circuitry, though, of the orange-sized human
equivalent, about realtime or 10-100x speedup).
> vinced that the difficulties of getting to trans-human intelligence
> are as you present them. I mean if the development of trans-human
> intelligence really is going to take:
>
> > truly impressive hardware, which will need oodles of cash and manpower
> > for R&D,
>
> Then maybe we should consider turning the job over from the computer
> science field to the gene splicers and the pharmaceutical industry; at
> least there, the starting platform is already one of "human intelligence".
I was assuming de novo protein design for structural matrix, de novo
enzymes, scaffolding, autoassembly, etc. Clearly silicon tech is dead on
the 20-30 decade range. They might migrate to molecular circuitry via the
Langmuir-Blodgett route, but I honestly don't think that is likely. It's
time to start afresh, substrate change being mandatory as we move up one
more dimension (2d->3d) and collapse structure size by one or two orders
of magnitude (can't say too much about power demands and switching speeds
yet).
> All in all, it seems like the greatest problems are not knowledge, but
> those amorphous things we call politics, wisdom, morality, motivation,
> and human behavior in general (not to mention trans-human behavior).
Exactly. Labs, lots of bright researchers, cash, time. A coordinated
world-wide programme to dwarf HUGO & Co. Do we have the resources? Do we
have the motivation? Even more importantly, do we have the time? There
are also a number of different, less pleasant developmental
discontinuities predicted for 2030, or so. Considering, how monumentally
stupid or species behaves on the large scale, I do not think them that
unlikely. We only move our ass when we feel the heat. This time, we
shouldn't make a mistake. We won't get a second chance.
ciao,
'gene
> --
> In the Ecstatic Service of Life -- Omega
>
>
______________________________________________________________________________
|mailto:ui22204@sunmail.lrz-muenchen.de |transhumanism >H, cryonics, |
|mailto:Eugene.Leitl@uni-muenchen.de |nanotechnology, etc. etc. |
|mailto:c438@org.chemie.uni-muenchen.de |"deus ex machina, v.0.0.alpha" |
|icbmto:N 48 10'07'' E 011 33'53'' |http://www.lrz-muenchen.de/~ui22204 |
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:44:04 MST