[extropy-chat] Video of Fighting Humanoid Robots!

[Adrian Tymes]

--- Emlyn <emlynoregan at gmail.com> wrote:
> ... apparently humanoid robots can do backflips and
> cartwheels. I had
> trouble believing my eyes. Incredible (and cool as
> f*?k).

A chain of thought I've been having, which this
reinforces:

Take basic robotics research, as applied to humanoid
frames.  Follow it on its current path until, for
whatever purpose (entertainment, military, etc.),
human-sized, human-shaped robots with
human-equivalent manipulation ability become
available.  The hardware is what's important here; the
software would, say, allow it to perform
pre-programmed dance steps or walk along a path as a
"human drone" for military training, but it would be
no more creative or free willed than today's robots.
Is there any logical doubt, given these toys being
available today, that this is likely within the next
20 years?  (Being conservative here - less than 10
might be more accurate.)

If not, then this frame will likely have some cargo
capacity somewhere within its body - probably its head
and/or torso.  And some extra power capacity, enough
to power the life support unit for a brain, and wire
it up to the sensors and motors, and possible hormone
synthesizers to replicate the missing organs (as part
of life support).  If there's a lot of cargo capacity,
more of the body can be encapsulated as a preliminary
measure, while researching how to support just the
brain (primarily, figuring out the hormone balance and
neuron wiring).  Is there any logical doubt that this
would be likely within 20 years of the above?  (Again,
being conservative.)

Once that happens, there exists a powerful symbol for
taking the last step, which is likely to focus and
drive AI research into uploads.  Which therefore makes
this seem the most likely "default" path to SIs, if
nothing is done to force another path before then.  As
nasty as human beings can be towards one another, few
- and none that are likely to gather the resources to
be among the first uploads - are as sociopathic as
"pure" AIs, which would have no pre-AI socialization,
are feared by some to be.  (One might take it a bit
further and wonder if research into "pure" AIs thus
makes anti-Friendly AIs actually more likely, but at
this early stage, that's way too difficult to
determine to be worth any action, or even much serious
thought...again, at this time.)