> Ellen Ullman, "Close to the Machine", (City Lights Books, 1997):
> What worried me, though, was that the failure of the global electronic
> system will not need anything so dramatic as an earthquake, as diabolical
> as a revolutionary. In fact, the failure will be built into the system in
> the normal course of things. A bug. Every system has a bug. The more
> complex the system, the more bugs. Transactions circling the earth, passing
> through the computer systems of tens or hundreds of corporate entities,
> thousands of network switches, millions of lines of code, trillions of
> integrated- circuit logic gates. Somewhere there is a fault. Sometime the
> fault will be activated. Now or next year, sooner or later, by design, by
> hack, or by onslaught of complexity. It doesn't matter. One day someone
> will install ten new lines of assembler code, and it will all come down.
*******************
http://csf.colorado.edu/pen-l/2000III/msg03330.html
Re: Contradictions abound
by Lisa & Ian Murray
22 September 2000 20:25 UTC
--------------------------------------------------------------------------------
>>Conveniently failing to notice that this same spread of real-time
information adds exponentially to the variables and the dynamic relations
between 'em all. I mean, Greenspan has a point if you define 'information'
as a 'lessening of uncertainty', but that'd mean you have to call stuff
that's (inter alia) wrong, polysemic, irrelevant, and decontextualised
something other than 'information'. Also, as more information becomes more
available, more information needs to be processed and cross-referenced,
because everybody else's state of information has been altered, too. And,
anyway, institutions are all about market share in these good times, no? I
mean, until things go pear-shaped, the bank manager is accountable to the
boss on the criterion of aggregate loans made, not how closely the loans
authorised approximated some degree-of-confidence calculus. After all, a
bank'd go bust being responsible in America right now ... else you'd neverf
be able to get stock speculation loans at 10% down, or margin-call deals on
VISA cards ... and I'm given to believe those are not at all hard to get
right now.
Cheers,
Rob.
========
G'day Rob,
You been reading Peter Albin lately? Or perhaps Joseph Tainter's ideas on
diminishing returns to complexification?
"I think recent work in computational complexity theory raises the
possibility that there may be another "critical mass" for a knowledge
representation, a maximum size threshold above which belief systems must in
effect disintegrate. For a representation to qualify as being understood by
an epistemic agent, the agent must be able to perceive an adequate
proportion of the interrelations among elements of a set. Otherwise, the
agent will not be a ble to identify and eliminate enough of the
inconsistencies that arise...The range of intractability results leads one
to wonder in turn whether knowledge systems of some finite size may be so
computationally unweildy in this way as to shatter...[Christopher Cherniak
"Minimal Rationality"]
At the same time we should avoid looking for contradictions behind every
bush....lest we metacontradict ourselves,
Ian
