Matt Mahoney wrote:
--- Richard Loosemore <[EMAIL PROTECTED]> wrote:
Hmmmm... I think my point may have gotten lost in the confusion here.
What I was trying to say was *suppose* I produced an AGI design that
used pretty much the same principles as those that operate in the human
cognitive system (non-determinism and all).
Under those circumstances, your test would fail to classify it as an AGI
even though it clearly would be an AGI.
Doesn't that make the test useless?
Only if you insist on using nondeterministic hardware. But why would you do
that?
Why do you ask? It makes no difference. The point is that your test is
meaningless because of the fact that it doesn't work for one class of
AGI systems.
As it happens, there are reasons for wanting to do it that way, and
there may well be reasons why the "deterministic" way of building an AGI
is for all practical purposes impossible. Under those circumstances,
all this talk of compression tests is just so much fantasy.
But, to reiterate, I don't need those further tests to make my case:
the mere fact that a perfectly functional AGI system would be classified
by your test as being not an AGI is enough to make the test a failure.
Richard Loosemore.
-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?member_id=231415&user_secret=fabd7936
