Meeting notice: 12-15-98; 7:30 p.m. at the Royal East
(782 Main St., Cambridge), a block down from the corner
of Main St. and Mass Ave.

Suggested topic: local autonomy/remote operation


While artificial intelligence and teleoperation are
often presented as alternatives they are obviously
symbiotic.  The natural model for reasonably advanced
machines is one in which local issues such as
orientation, obstacle avoidance, and manipulator
control are handled locally, autonomously, while humans
feed the machine directions on high-level matters such
as targets, output specs, and rates.

While at present this model is usually found only in
machines built to work in dangerous environments, it is
easy to imagine it in a wide range of contexts,
including agriculture, construction, transportation,
mining, and manufacturing, not to mention a host of
research applications.  Of course nanotechnology is
entirely inconceivable without a comprehensive
development of this model.

It is worth noting that the humans issuing these
more general directions are likely to be physically
separate from the machines themselves, for reasons that
include safety, economy (accommodating human operators
inside or on the machine itself adds expense and often
degrades function), and efficiency: local autonomy
permits operators to run large numbers of machines in
many locations simultanously.

One implication of remote operation is how it defines
the path of sensor and AI development: once the market
becomes familiar with the core idea it will be simple
to pitch successive generations of products, each
promising incremental increases in the number of
machines that can be controlled by a single human
during a specific productive cycle.  Perhaps there is a
Moore's law type curve in here somewhere.

There are also intriguing cultural implications.  Many
cultural roles depend on a close physical association
with machines.  There are many examples -- truckers
wrestling with their rigs, pilots, skippers -- but the
most extreme example is the connection we have with our
guns.  Police work might be very different, both in how
it feels as a job and in its cultural resonance, if
guns were controlled by patrolmen in an armored car
telling a fleet of microhornets to return fire, or by
guys sitting in cubicles downtown clicking on spots on
a screen. Local autonomy also destroys the exclusivity
of access to complicated machines, both by lowering
operating skill requirements and by opening up the
possibilities of remote audiences.

Finally, local autonomy and remote operation make it
unnecessary for anyone to expose themselves to
dangerous situations of any kind.  Once this fact is
internalized, movie makers interested in showing heroes
going into danger will be forced to rely on historical
(or perhaps alternate universe) dramas, since most
audiences watching a contemporary hero take chances
will wonder why he didn't do what anyone else would
have: send a machine in instead.


Past business:

William Ware <wware@xionics.com> reports the discovery
of a web page summarizing George Whitesides' talk on
soft lithography:
http://itri.loyola.edu/nano/us_r_n_d/04_02.htm


Nanonews:

Microlab slashes cost of DNA testing.
http://www.umich.edu/~newsinfo/Releases/1998/Oct98/r102198b.html

Rice University and NASA officials form nanotube
research consortium.
http://www.eurekalert.org/releases/ru-nmdnff.html

cybernetic materials
http://www.eurekalert.org/releases/im-sms.html

enzyme engineering
http://www.eurekalert.org/releases/ciw_cdpe.html

neural materials (resonant tunneling diodes)
http://www.eurekalert.org/releases/ns-sssco.html

Recombinant biochem

Adapted from the release: A current method of making
new proteins is to use ultraviolet light or a
DNA-disrupting chemical to make genes mutate and make
proteins from the mutations.  Typically only 1 in a
hundred mutations made this way is interesting. By
comparison, in the reported test 13 per cent of the
enzymes produced through "DNA shuffling" were superior
to the original. The best, a version of interferon,
was 285,000 times as potent.
http://www.eurekalert.org/releases/ns-wgbitl.html

Smallest possible bipolar transistor is built.
http://www.eurekalert.org/releases/weiz-hmcwg.html

Announcement Archive:
http://world.std.com/~fhapgood/nsgpage.html.


Comments to: hapgood@pobox.com