From: Dani Eder (firstname.lastname@example.org)
Date: Thu Dec 09 2004 - 14:57:30 MST
> My reasoning is this: sometime between 2020-2030,
> nano revolution is going to start hitting, and the
> risk from existential threats (Terrorist goo,
> Unfriendly A.I) is really going to start becoming
Manufacturing productivity trends point to highly
automated production arriving in 10-15 years. The
trends I track are for the US durable goods sector
(things designed to last more than 3 years), which
is the total productivity for all existing plants
in place no matter how old or new. The trend for
new factories is already highly automated. For
example, a current news story:
says Samsung will invest $24 billion in chip
factories in the next 6 years, which will yield
10,000 new jobs. That's $2.4 million in factory
investment per job.
Compare that to an established
set of factories like Boeing's (whom I work for).
We have $8.3 billion in plant and equipment for
157,000 employees, or $53K per person.
It's just going to take a while for the new highly
automated factories to take over the bulk of the
Once retail stores like WalMart go to robots for
stocking shelves, and automated self-checkout, and
factory production goes highly automated, there are
two big issues:
1) What happens to the ~1/2 of the workforce that
has had their jobs automated?
2) What becomes possible when a set of automated
factories can produce all of each other's parts,
or enough of each other's parts plus selling
surplus production to pay for the remainder, so that
they become a self-sustaining ecosystem?
These issues will be hitting before the nanotech
and AI risks, most likely. The manufacturing jobs
issue is already politically significant in US states
that have a high share of such. In this
economic recovery, productivity improvements are
eliminating jobs faster than demand increases are
creating them, so there is a net job loss in the
> A person 40 or under today has a greater than 50%
> chance of living to 2040 naturally, even assuming no
> significant life extension. So you see, the
> probability of an under-40 dying from existential
> threats assuming no FAI is roughly comparable to the
> probability of them dying through natural wear and
> tear. The rational option is to prevent death then
My plan involves building a strong defensible
residence in a remote location in case of civil
disruption in the next 10-15 years. It happens it
will look like a castle, because medieval re-enactment
is my hobby. But castles were defensible residences
in their time. It won't stop a government attack,
but it should be proof against looters with small
Part of what I want to do at the residence is study
The power source problem looks already solved.
Spectrolab, a division of Boeing that makes solar
cells mostly for satellites, has a triple-junction
cell that is 30% efficient at 400 suns. So while
the cells themselves are more expensive per sq cm
than conventional solar cell panels, you don't need
very many square cm of cells, and reflectors are
much cheaper per square meter than semiconductors.
Overall they require 1/10 the dollar value of flat
panel cells for the same output, which makes solar a
cost competitive power source.
> The one thing I'm reasonably sure about is this: if
> we don't get FAI before 2045 or so, then, to borrow
> phrase from Damian Broderick's 'The Spike': '"We
> all perish most horribly"
What we need is for life expectancy to improve ten
times faster than it has been (i.e. 1.0 year per
year rather than 0.1 year per year). At the current
rate of improvement, you can expect to live a few
years longer than the previous generation. At a
sufficiently high rate of improvement, your chances
of lasting long enough for a long term-solution like
uploading improve dramatically.
That can come in theory from piecemeal medical
advances each of which only adds a few years, but
in the aggregate add decades.
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
This archive was generated by hypermail 2.1.5 : Sat May 18 2013 - 04:00:45 MDT