Re: MEDIA: Forbes Magazine discusses AI and the Singularity (well, kinda)

Michael Nielsen (mnielsen@tangelo.phys.unm.edu)
Thu, 25 Jun 1998 14:30:24 -0600 (MDT)


Message reformatted, since the paragraphs show up without line breaks
on my machine.

On Thu, 25 Jun 1998, Doug Bailey wrote:

> "Kurzweil's long-term impact, however, is much greater than just speech
> recognition. His belief in the exponential growth of technology gives
> voice to humanity's possibilities. He is a visionary whose predictions
> are based on hard science. [Doug note: Last I checked, Grove's Law runs
> into a problem called "physics" around 2012 or so. Kurzweil does not
> offer any insight into how this problem might be overcome - though
> parallel processing might work. ]

The curves I've seen make 2020 look like the ultimate limit, for reasons
of both size and heat dissipation. Various means for circumventing the
obvious physical limits may exist though.

I would be interested to hear what people think about the possibility of
extending the reign of Moore's law beyond 2020, when we hit atomic-sized
components, dissipating kT Joules per operation, assuming it continues on
its merry path for the next 20 years.

What are the possibilities for increasing computer power beyond that point?

> 200 mips reached in 1998: Speech recognition achieved [Doug note: Didn't
> Univ. of Tokyo's GRAPE-4 exceed 1,000 mips already?]

(Do you mean bips, not mips?)

ASCI Red at Sandia Lab has reached up around 1,830,000 megaflops.

Michael Nielsen

http://wwwcas.phys.unm.edu/~mnielsen/index.html