Re: The Spike, nanotech, and a future scenario

Brian Atkins (brian@posthuman.com)
Sun, 05 Oct 1997 20:54:29 -0400


Hal Finney wrote:
>
> Eliezer S. Yudkowsky writes:
> > More in the sense of potential than probability. I think that the most
> > probable time is actually 2008... using my oddest calculating method yet. I
> > figure that in the fifteen years between 1980 and 1995, the E.T.S. went from
> > 2035 to 2020. That is, each year the projection moves up by one year. If
> > this continues, the Singularity should occur in 2008. I find this argument to
> > be strangely compelling, perhaps because it sounds so meta.
>
> Actually I find that the Singularity stays about 20-30 years in the
> future. I'd like to see some evidence that anyone in 1980 predicted it
> to be in 2035, 55 years in their future. I don't think anyone even knew
> about the Singularity back then.
>
> 20-30 years is enough time for significant changes in technology and
> in society. When the movie 2001 was made, it did not seem implausible
> that a space station, lunar bases, AI, and a manned mission to Jupiter
> could occur in that time frame. In fact, technology has not advanced
> as fast as was expected back then, at least not in those areas. AI has
> been 20-30 years away since the 1950's.

This is a good point, and I think that visions of a Singularity
based on nanotech suffer from this. However...

>
> Personally, I think we are going to hit a big wall in computer technology
> in the next ten years. They can't keep making silicon features smaller
> indefinitely, the electric fields (voltage over distance, and voltage
> can't drop below the diode bias voltage which is an inherent aspect of
> the chemistry) if nothing else will cause problems. Other technologies
> will have to be developed to replace silicon. It may happen, but there
> is nothing which looks very practical as a replacement right now.
>

I have to disagree here. Things like x-ray lithography and copper
chips recently demonstrated can easily keep the industry afloat
for the next 10 years. Even after that there are candidates
currently being successfully researched like: optical logic gates,
quantum computers, quantum dots, and other stuff I don't know
because I'm a software guy not hardware... All of these advanced
pieces of technology have been demonstrated to actually work,
although they are very far from being powerful enough to even
replace a Pentium. But at least there are possibilities, and work
being done on these in the next 10 years should provide ways to
keep Moore's law going.

> It's our old friend, the S curve. Right now the growth in performance
> looks exponential, just as the increase in sewing machine speeds (or
> virtually any other industrial performance measure) looked exponential
> at one time. But they hit a limit back then, and we will probably hit
> a limit ourselves now.
>

I'm relatively new to the list, but my understanding of an "S curve"
is that is for a particular technology- like aluminum+silicon
chips like a Pentium. Obviously each technology reaches a limit,
and something new like copper steps in. The question is can we
keep finding newer and more powerful technologies, right?

> That doesn't mean that progress will stop; all the ramifications of
> the information revolution will continue to develop even if computer
> performance tops out a few orders of magnitude better than we have
> it today. And there are other technologies which are poised for growth,
> biotech of course, and possibly materials science, microtech, etc. These
> are probably going to be the hot fields of 2010.

What is interesting is that all of those fields relate back to
computing and may make more powerful computation available...

>
> But I am always suspicious of that 20-year prediction horizon. We can
> guess what will happen technologically in the next ten years, but beyond
> 20 we really have no idea. "Here there be dragons," and we are inclined
> to put our wonders safely in the 20-30 year period. In practice though
> things often take much longer than we expect.
>
> Hal

For anything related to software engineering (AI), I agree with
you. Unless we can come up with ways to reduce the complexity of
software development, things there will get slower and slower.
Hardware on the other hand should continue to grow for at least
another 20 years even with the technologies we have today. The
question is: does super-hardware without AI software bring about
a singularity? Or will the super-hardware make the effort of
programming an AI possible?

-- 
The future has arrived; it's just not evenly distributed.
                                                       -William Gibson