> Samantha Atkins wrote:
> > Hardware is only part of the story. Where is the software? Who is
> > developing it? On what schedule? With what deliverables allowing
> Um, do you see any breathtaking progress in software productivity,
> plus a silver bullet to the complexity threshold emerging within
> the next few decades, outside of evolutionary algorithm field?
> If yes, I'd like to see pointers, because I'm apparently missing
> that hot new stuff fermenting somewhere.
Well no, at the moment I don't see the breakthrough. As a matter of
fact, from my couple of decades inside the software industry I see a lot
of forces arrayed against any considerable efforts towards such a
breakthrough being made. Hell, let's be blunt. There is a lot of
energy and momentum arrayed against even incorporating real improvements
to software design and development made over a decade ago! Which help
motivate the question. Evolutionary algorithms are NOT a silver bullet
for the problem. Far from it.
> I wish software people would stop seeing the world as made of
> nails. The neuronal/numerical modelers/ALifers are those most
> likely to suceed, not the Comp.Sci./AI people. Unless they can
> whip up a couple of new paradigms, pronto.
Everyone has their prejudices. I don't think either of the groups
mentioned have what will be needed, yet. I don't think neuronal models
and ALife stuff will give us the breakthrough we need although it will
give many valuable tools and critical understanding of how minds such as
ours work. Nor would I write off Comp-Sci and AI so quickly. We need
everything we've got and then something we only begin to see or know how
to do yet.
> > tracking of progress?
> I wonder why you allocate any of your attention to this particular
> project. It is not even vapourware yet.
True on the one hand. On the other hand, it is seen as urgent enough
and as salvation yet that I would like to see it start moving as a
project - just on the off chance that it is the only hope after all.
Although frankly I think the notion that a "friendly" SI will solve all
of our most critical problems is extremely unlikely.
This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:56:21 MDT