[ExI] singularity thru simulation of evolution
nanite1018 at gmail.com
Mon Nov 16 04:02:19 UTC 2009
On Nov 15, 2009, at 6:20 PM, spike wrote:
> What I do mean is this: what is your mental picture or roadmap,
> fuzzy, to the singularity?
> ......Is there an alternate path to the singularity that does not
> require a deep
> understanding of how intelligence came about to start with? Is
> there any
> path to the singularity that does not require an understanding deep
> to create a simulation which successfully demonstrates the evolution
> intelligence? If so, do share.
While I don't believe in some kind of Singularity (with a capital
"s"), I do think that there are going to be a series of major changes
over the next hundred years, and I think any one of them may deserve
to be called singularities of some form or another.
My singularities are:
1. Radical life extension and the achievement of so-called longevity
or actuarial escape velocity (just for clarity, when we increase life
expectancy by one year every year). This will totally transform human
civilization. The path to get there is through developments in
biotechnology and eventually nanotechnology. Of course, mind uploading
would also be a good way to get to this point, but I think that will
come along after we achieve biological indefinite lifespan.
2. Nanotechnology, specifically molecular manufacturing along
Drexlerian lines, either through assemblers or nanofactories. This
will fundamentally alter our economic situation, allowing for the
effective end of scarcity and poverty on Earth, even without advances
in AI. Will also be immensely helpful for development and growth of
human-level and above AI, and for the development of mind uploading.
3. The development of human level AI. This will change our society by
challenging its laws, transforming its economy, and affecting our
culture. This could easily fuel either of the two other singularities,
or either of them may help fuel this one.
I think the likely path to the singularity is the development of
advances in biotech, which will extend longevity enough to where we
get, say, halfway to LEV. At that point, all those Baby Boomers will
invest absurdly large amounts of money into life extension research,
including large amounts into nanotech, which will push us rapidly
toward LEV. Around the same time, advances in nanotechnology will
enable major medical advances, allowing us to achieve LEV, at least
for the world's wealthy (e.g. middle class Americans and up).
The money from longevity research will round out any problems
nanotechnologists haven't been able to solve with molecular
manufacturing (since it will likely be necessary for the mass-
production of the more rigorous and effective longevity and medical
techniques and technologies).
And also around this time, the advancement of computer power and
neuroscience will allow us to use the nanobots and the like we are
using to extend our lives to also investigate the brain, allowing us
to put the finishing touches on our computer models of brain function,
and molecular manufacturing will allow us to mass produce our most
advanced computer systems with little input cost, ushering in a new
era of indefinite lifespan, scarce scarcity, and human-level and above
artificial intelligence (soon followed by mind uploading). On time
scale, I'd say that we reach halfway to LEV by 2030 at the latest, and
put the finishing touches on our multiple singularities by 2060 or so
(around when high-power computers are currently expected to be able to
handle the most extravagant estimates for processing power needed to
emulate the brain.
In regards to your evolutionary approach spike, I'm not sure its
necessary at all. We don't even need to understand intelligence or
consciousness in order to recreate it. All that's necessary is to
emulate the actual physical processes going on in the brain. That will
likely come along far sooner than a full and complete description of
the computational processes occurring in the brain, and the
foundations of consciousness. Evolution is more of an explanatory
theory than a predictive one (even if we know the end result), and
trying to postdict the evolution of organisms in the past is likely to
prove an enormously more difficult task than the above "singularities."
nanite1018 at gmail.com
More information about the extropy-chat