[ExI] AI Motivation revisited

Stefano Vaj stefano.vaj at gmail.com
Wed Jun 29 16:07:02 UTC 2011


On 27 June 2011 23:28, Richard Loosemore <rpwl at lightlink.com> wrote:
> Meaningless statement.  A trivial use of the PCE.
>
> An ordinary PC, if programmed correctly, might be capable of something
> approaching full human intelligence in real time.  Probably not, but it is a
> possibility.
>
> Since everything depends on the type of computation required (which you do
> not know), it is pointless to make statements such as "a contemporary PC is
> certainly capable of running an anthropomorphic AGI".

Let us make it easier, so that we can see whether and where a
disagreement exists:

- Beyond the level of universal computation, there is no other
"threshold" which may be required for any device to perform any given
computation, the difference obviously being in the relative
performance of different devices engaged in different tasks;

- An anthropomorphic AGI is a computation process of one kind or another;

- A contemporary PC is a universal computation device.

Hence, a contemporary PC, exactly as a Turing machine or a cellular
automaton of the right kind, is in principle capable of running an
AGI.

Fact.

Educated guess, but by all means not a fact:

Contemporary PCs seems unlikely to perform very well in any such task,
including in comparison with human brains, so that you for a
persuasive AGI might have to wait for aeons to get a reply in a
Turing-test interaction.

-- 
Stefano Vaj




More information about the extropy-chat mailing list