--- Eugen Leitl <[EMAIL PROTECTED]> wrote:

> On Sun, Jul 01, 2007 at 12:45:20AM -0700, Tom McCabe
> wrote:
> 
> > Do you have any actual evidence for this? History
> has
> > shown that numbers made up on the spot with no
> > experimental verification whatsoever don't work
> well.
> 
> You need 10^17 bits and 10^23 ops to more or less
> accurately
> represent and track what the human brain does.

So what? We're not going to build an AGI by modeling
every atom in the human brain. This is therefore an
upper bound on the computing power we need, because we
know we can get at least one kind of intelligence
(humans) using only this much computing power.

> We don't know whether above number is too low (it is
> probably high,
> and maybe even ridiculously high) but lacking other
> means of
> validation (with the exception of the retina) it's
> as good an
> estimate as any.

An upper bound is not a good estimate, especially when
the stakes are so high. The more important everything
becomes, the more conservative you want to be, and the
conservative assumption is that we already have more
than enough computing power on your average PC to run
an AGI. This is admittedly unlikely, but it's a good
estimate because it's less likely to lead to the Earth
getting destroyed than other estimates.

> Notice that you may need way more than that to
> bootstrap an
> intelligence by evolutionary means.

Why would you *want* to create an intelligence by
evolutionary means, even assuming that you could?

> -----
> 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/?&;
> 

 - Tom


       
____________________________________________________________________________________
Sick sense of humor? Visit Yahoo! TV's 
Comedy with an Edge to see what's on, when. 
http://tv.yahoo.com/collections/222

-----
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=4007604&user_secret=7d7fb4d8

Reply via email to