Re: Development models.

From: Michael Wilson (
Date: Wed Sep 07 2005 - 08:02:15 MDT

Michael Wilson wrote:
>> I would not be surprised at all if the first AGI is effectively the
>> product of 'one mad genius' architect, though that architect may
>> have a team of implementers following their instructions and
>> designing non-core support systems.

Note that when I say 'I would not be surprised', I mean 'I assign
this outcome a nontrivial probability', not 'I think this is the
most likely outcome' or 'I think this is how we should do it'.

Justin Corwin wrote:
> But it is, insofar as I'm aware, a theory without any evidence. I
> can't think of any one person who can seriously be credited not only
> with innovative theories, but a complex final design of a nontrivial
> part of our world. Things are much too complicated for that.
> Especially software design, wherein the interactions (even inside a
> system that could be said to be designed by one person) are often
> entirely unexpected to the designer.

This is at odds with both my experience of software engineering and
more importantly my model of what is necessary and sufficient for
seed AI. In particular you appear to be using 'large commercial
software engineering project' as the basis for your model of
'successful AGI development project', wheras I would now say that
while AGI development requires software engineering expertise it
doesn't conform to the software engineering project model. Assuming
that you hold this opinion for the same reasons that I used to hold
it, I doubt we're going to come to any agreement on this point.

Ben Goertzel wrote:
> Given the current state of hardware/software tech, it is not
> feasible for one person to put together a seed AI.

Frankly I wish that were true, as it would remove one-man projects
from consideration as existential risks, but unfortunately while
the probability of a single person doing everything is low it is
still nontrivial when aggregated over all lone researchers.
Clearly some people on this list (who are actually trying to build
AGIs on their own) would assign a rather higher probability to
this than I would. Again the discrepancy here is caused by a
yawning gulf between our models of what kinds of seed AI could
be built, and I don't expect a useful debate on this point when
there is no consensus on the prerequisite assumptions.

 * Michael Wilson

Yahoo! Messenger - NEW crystal clear PC to PC calling worldwide with voicemail

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:52 MDT