Matt Mahoney wrote:
On Fri, Mar 8, 2013 at 12:11 AM, Alan Grimes <[email protected]> wrote:
Why are you so fixated with this 10^10 number???
How about you worry about 10^0 or 10^1 first???
It is the population of the world, or more precisely, the working
population, that is of interest if we wish to replace all human labor
with machines. AGI would do that, right?
Creating a single human level AGI is certainly less expensive. I
estimate something like 100M lines of code on a 10 petaflop, 1
petabyte machine. But what would you do with it?
Make 10^n copies. ;)
You gotta learn to walk before you can run.
Before you can have 10^10 AI's, you have to have 10^9 AIs, and before
that 10^8 AIs, etc.
Most everyone wants to work on the software. For a single AGI, it's
the most expensive part. Then once it's written, you can make
unlimited copies. You'll be an instant quadrillionaire, except for the
small detail that supercomputers aren't cheap. What we really need is
a fundamentally new hardware design. Silicon won't work. You can't
make transistors smaller than atoms, which is what you would need to
do to reduce the power consumption to the 25 watts that your brain
uses.
I think you are being pessimistic about your hardware estimates, but
fine. Let's talk about
hardware then. There are two basic approaches, 3D nano-crystals and
quantum computing. You can also jack up performance, by a factor of 100,
using custom designed chips.
--
E T F
N H E
D E D
Powers are not rights.
-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription:
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com
