Re: Singularity and AIs (was: We're stuck with each other)

From: Robert J. Bradbury (
Date: Sun Jan 27 2002 - 15:16:36 MST

On Sun, 27 Jan 2002, animated silicon love doll wrote:

> 2002.01.27 3:36:22, Eugene Leitl <> wrote:
> >So there is a very real hardware problem, namely that 2 GOPS of braindead
> some things here go over my head,

Chesire -- I am happy to agree with you on this point. Particularly
with regard to Eugene's comments. Most of the time he gets my vote
as the most obscure, condensed poster on the list (of course I can't
view myself clearly -- the recent exchange with Louis suggests that
I could very well fall into that swamp myself). I am generally
left looking at his comments questioning whether I really understand
them -- and most of the time I think our relative knowledge bases are
pretty similar -- which makes my confusion sometimes even stranger.
I would write it off to cultural differences but I of all people on
the list have at least some overlap with Eugene's cultural background.
So the source of this remains a mystery to me. So viewing Eugene's
comments as an enigma wrapped inside a riddle boxed within a puzzle
is not a new situation.

> but until now i don't think i've seen anything that i just completely
> don't get. so. what's a GOP?

That's easy. G = giga or "billion". OPS = "operations per second".

To hang tight on the extropians list you have to learn the acronyms.
Yo gotta no gigas and teras and petas and exas and zettas and yottas
and yoctos and zeptos and attos and femtos and picos and of course
most important of all the nanos.

> slightly less important, but on a similar note, could anyone give me
> a precise definition of computronium?

Its a catch-all phrase for matter structured in a way that allows
computation. In my experience, Robert Freitas may have invented
it. It might have come up on the Extropian list before I first
heard it from him (I've never checked this). Because different
computational architectures are optimal for different types of
computations (neural networks vs. cellular automata vs. quantum
qubits come to mind) there doesn't seem to be a way to specify
the "best" way to arrange matter to produce results in the
fastest way. So "advanced intelligences" would first decide
"what" they want to think about, then they would decide the
"optimal" hardware architecture that can be used to think about
that, then they would construct (or evolve) that architecture
and build the "computronium" and finally do the thinking.

In our current reality it would be the difference between the
Thinking Machines (now defunct) "connection machines", the
GRAPE computers for gravity computations, IBM's "Blue Gene"
processor-in-memory pseudo-cellular-automata architecture
for protein folding (and possibly other applications) and
"general purpose" von Neumann/Harvard architectures (99.99...%
of what is currently available.)

All of those are examples of "computronium". Each architectures
has applications for which it may be best suited.

> we all create life, some of us with wombs,
> some with smiles, some with patient hands.
> we are all gods, if we choose to be.

I like to create mine with a Genome Editor and a DNA synthesizer.


This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 13:37:36 MST