From: Jef Allbright (firstname.lastname@example.org)
Date: Thu Feb 23 2006 - 13:35:14 MST
On 2/23/06, Richard Loosemore <email@example.com> wrote:
> When Kennedy gave his famous speech at Rice University about going to
> the Moon, the speech, as well as the other things he said to his
> subordinates on the subject in the next few months, could probably have
> been contained in some 30K bits. Maybe? Ballpark, anyhow.
> But when this small packet of information was launched into the society
> that was America, it triggered some massive shifts of resources that
> eventually resulted in a series of Apollo spacecraft landing on the Moon.
> It tells us almost nothing about the Apollo space programme to speculate
> on how many bits were in JFK's utterances.
> Ditto for the amount of information in the DNA that triggers the
> development of a brain.
You make a valid point, within the context you intended.
However, within a broader context of discussion, at least two
additional points do appear to be significant:
(1) Any AI capable of interacting effectively with our world would not
be not fully defined by its own internal code, and especially not its
seed code, just as any human is not fully defined by its genetic
legacy. Development plays a very large role, and is highly dependant
on the environment. In a very real sense, there is "intelligence"
encoded in the environment as a result of long evolutionary processes,
and this external "intelligence" plays a large role in what we
perceive as intelligent behavior exhibited by what we perceive as an
I have expanded on this in the past, talking about the illusion of a
discrete Self, but I appreciate that such topics can be seen as "mere
philosophy" and not applicable to practical AI.
(2) I do think that a valid comparison can be made between the the
algorithmic complexity of our genetic code and the algorithmic
complexity of seed AI code, with the understanding stated in (1)
above, that development within a given enviroment plays a large role
in what we come to see as intelligent (and largely heuristic) behavior
within that environment.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT