Re: Moravec's estimates?

From: James Higgins (jameshiggins@earthlink.net)
Date: Sat Apr 07 2001 - 11:16:13 MDT


At 05:56 PM 4/6/2001 -0700, you wrote:
> >
> > You have to say *why*, or it doesn't count.
> >
> > -- -- -- -- --
> > Eliezer S. Yudkowsky http://intelligence.org/
> > Research Fellow, Singularity Institute for Artificial Intelligence
> >
>
>Well, any prediction on when any x-flops machine would be out has been
>beaten so far. Weren't thoses predictions based on doubling power every 18
>months ? I might be mistaken...

I'm sorry, but I don't follow. Are you saying that it will be available
sooner or later or ?

Moors law has held quite well for roughly 2 decades now. So I'm inclined
to believe it will probably remain so for the near future (decade or so).

However, their may be one major problem with this that could put a halt on
the whole AI/Uploading picture. Moors law will only continue as long as it
makes economic sense for the companies involved (Intel, AMD, etc). At the
moment MHz is what sell's processors, but their are no killer apps to use
the speed that is now available. If the computing power continues to
outstrip demand, we may see a sudden halt to faster/better/cheaper
computers because of falling profits.



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:36 MDT