[extropy-chat] Reading "The Singularity is Near"

Dirk Bruere dirk.bruere at gmail.com
Thu Sep 22 21:43:15 UTC 2005


On 9/22/05, Mike Lorrey <mlorrey at yahoo.com> wrote:
>
>
>
> --- Marc Geddes <m_j_geddes at yahoo.com.au> wrote:
> >
> > The secret I think is this: A general recursively self-improving
> > friendly AGI is an AI which understands the *true* nature of mind.
> > So to initiate Singularity, we simply design the AGI to understand
> > the *true* nature of mind and the rest (friendliness etc) follows
> > automatically as secondary consequences.
>
> On the contrary, IMHO such a desired AI should instead understand that
> it is a flawed, imperfect being, a unique, freak, and yet happy
> accident that cannot, no matter how intelligent it is, consider itself
> the be-all, end-all. As it is the creation of flawed beings, so too is
> it flawed and must focus on its own improvement with acceptance that
> perfection is likely unattainable, and thus perfect understanding of
> the true nature of mind is similarly unattainable.
>

IMO the first true AI is likely to result from a kind of uploading where
someone's brain is sliced down to the molecular level, scanned, input and
run in simulation. Since petaflop computers are likely to be around from
about 2012 onwards I'd say that pre-2020 would be a likely timeframe.

Dirk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20050922/5a1cf375/attachment.html>


More information about the extropy-chat mailing list