[extropy-chat] "The Spike" - Raymond Kurzweil

[Damien Broderick]

----- Original Message -----
From: "Robert J. Bradbury" <bradbury at aeiveos.com>
Sent: Wednesday, October 29, 2003 7:43 PM

> > But I considered his
> > own special contribution, the `Law of Accelerating Returns',

> > http://www.kurzweilai.net/articles/art0134.html?m=1

> > to be at the verge of bogosity (sorry, Ray) when regarded as a sort of
> > cosmic Truth.

[Robert comments:]
> Ray is pretty good compared with others (Ron Klatz of the A4M
> comes to mind) when it comes to pushing trends (e.g. the doubling
> rate for medical knowledge) way beyond what is reasonable.
>
> Now, the "Law of Accelerating Returns" is an interesting concept.
> For example, clearly the Google Search of Books will accelerate
> progress if it survives the assault by the authors guild. [etc]

I was going to wait until Pietro had told us which of Ray's ideas he wanted
to see analyzed in depth. But I should probably nip this line of objection
in the bud:

My incautious labeling of some of Dr Kurzweil's ideas as verging on the
bogus comes only (as I say above) from the *cosmic* or *general* nature of
his proposed `Law'.

*Obviously* computer bang for buck, etc, has been increasing pretty much as
Ray charts so usefully in the essay cited. This is the very basis of the
technological singularity discerned by Dr Vinge and discussed in my book and
Ray's. The question is whether this contingent fact of local technological
history can be projected backwards across space and time to yield a general
`Law of Accelerating Returns'. To cite that paper:

`Evolution applies positive feedback in that the more capable methods
resulting from one stage of evolutionary progress are used to create the
next stage. As a result, the rate of progress of an evolutionary process
increases exponentially over time. Over time, the "order" of the information
embedded in the evolutionary process (i.e., the measure of how well the
information fits a purpose, which in evolution is survival) increases.

`A correlate of the above observation is that the "returns" of an evolution
ary process (e.g., the speed, cost-effectiveness, or overall "power" of a
process) increase exponentially over time.

[...]
`Biological evolution is one such evolutionary process.'

Now this *might* be true (although bacteria don't seem to have `progressed'
yet they're doing just fine), but it seems to suffer from the hazards that
Karl Popper tore at in his great polemical book THE POVERTY OF HISTORICISM.
It reifies `progress', it assumes that there is a measure of progress that
remains identifiable over cosmic or at least geo-biological time scales.
This seems to be a remnant of 19th century thinking that's very pervasive
among technophiles but extremely difficult to defend.

What's more, it makes no allowance for the co-determination of evolved
structure and context, let alone the contingencies (such as random
intrusions like asteroid strikes or GRBs) that might produce the gross
discontinuities of punctuated equilibrium.

None of this means that Ray's *forecasts* concerning AI are dubious or even
implausible. I find them compelling, and was persuaded by the same factors
long before I read his work. It *does* mean that he's chosen (IMHO) an
unfortunate metaphysical and rhetorical framework for the exposition of
these important ideas.

Of course, on the other hand, in the world of public discourse this might be
a clever move--akin to the playful cavortings and winkings of Brian Greene
in his Elegant Universe series on string theory running right now on TV--but
it could serve to deter serious scholars who regard this framework as a
version of long-exploded teleology.

But this is such an important issue that I really *do* hope some
philosophically astute people here will throw in their 2 cents' worth. Max?
Greg? Anders? Anyone?

Damien Broderick