From: Philip Goetz (philgoetz@gmail.com)
Date: Sun Apr 23 2006 - 11:06:34 MDT
On 4/19/06, Ben Goertzel <ben@goertzel.org> wrote:
> The argument for the Singularity is that more radical technologies
> like AGI, and powerful nanotech and genetic engineering are on their
> way. The exponential curves, and the power of PC's and the Internet,
> are relevant to the Singularity only as part of the overall
> qualitative argument. What Kurzweil did in his book is to lay out the
> qualitative argument in a straightforward and digestible way. I think
> that in his presentation, Kurzweil over-hypes the exponential curves
> aspect, but even so, he does present the qualitative argument quite
> competently.
>
> -- Ben
I think this is the critical point (no pun intended) in this argument.
We should distinguish between 2 types of Singularity.
The classic Singularity, as popularized by Vinge and now Kurzweil (I
think Ben G. said that Von Neumann used the term in the same way?), is
that the Singularity is a divide by zero on the timeline, the place at
which you can't calculate the summed change because your calculations
go to infinity. This relies on exponential curves.
The Singularity that Ben is talking about is linked to specific
technologies. As anticipated by several people on this list, it is
linked most specifically to AI. The notion is that once we reach
smarter-than-human AI, it will bootstrap itself, making exponential
gains, so that within a matter of days life as we know it could
disappear.
I don't think this is inevitable, because I think AI will gradually
accumulate domains in which it is smarter than humans, rather than
crossing the border all at once. Nonetheless this "all at once"
notion is what we need a new name for. It is distinct from the
classic Singularity in several ways:
- It doesn't depend on whether technological change is increasing or
decreasing at the time it happens.
- It doesn't provide any acclimatization period during which people
get used to the exponential curve.
- It doesn't provide any warning.
It isn't a smooth curve. On the timeline, it's more like a wall.
Technological change may increase or decrease; it really doesn't
matter - whatever rate of change humans achieve won't prepare them for
the day Humanity hits the Wall.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT