Samantha Atkins wrote:
> Because time and therefore processing and therefore certainly
> learning will not proceed instantaneously immediately upon
> Singularity occurring. Processing power can only be grown at
> physically possible speeds. Using that capacity fully and well
> proceeds even more slowly. Learning I expect proceeds more
> slowly still. I think it is a mistake to assume that all that
> is knowable will be known almost immediately after reaching
I didn't say "immediately". I said "thirty physical seconds". That's
rather a lot of time.
> There is a cartoon from the series that I think was called "Bob
> the Angry Flower" that shows instead an angry Galaxy Brain.
Yes, this was the cartoon that was forwarded to me by a friend (Patrick
McCuller) and that turned me into a regular reader of BATF. I once
emailed Stephen Notley (the cartoon writer) asking if I could put that
image (or rather, a piece of it) on a single T-Shirt. Later events
obviated this and that image has yet to make it to a T-Shirt, as far as I
know, but at any rate I am quite familiar with the image to which you
refer and once played with it in Photoshop.
"Bob the Angry Robo Galactic Mega Brain"
> the cartoon the galaxy sized mind is complaining to others of
> its kind that it has computed all possible things and finished
> so many picoseconds ago and now there is nothing more to do
> forever! A small cluster brain
That's Freddy the Flying Fetus.
> responds, "I told you not to
> compute It all at once. I told you you would be sorry!" :-)
Yes, that's the pessimistic formulation of Eternal Fun Theory. Under the
optimistic formulation of EFT, a brain the size of the galaxy would
effectively never run out of Fun, even computing at Planck-time speeds
over a thousand times the age of the Universe.
> I don't think arriving at the point I spoke of is the Grand
> Finale at all.
Okay, just checking.
-- -- -- -- --
Eliezer S. Yudkowsky http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:39:58 MDT