--- On Fri, 10/24/08, Russell Wallace <[EMAIL PROTECTED]> wrote:

> http://wiki.alu.org/Russell_Wallace%27s_Road_to_Lisp

I think choosing a programming language for AGI is a bit premature. The purpose 
of AGI is to satisfy the goals of people. That role is currently served by the 
global economy, which has an algorithmic complexity of 10^17 to 10^18 bits. 
Almost all of that knowledge is stored in human brains. The maximum rate at 
which it can be extracted is 2 bits per second per person. That knowledge is 
not going to be coded in any programming language. In fact, it is far too 
expensive even to code it as explicit instruction in natural language. The 
cheapest way to extract it (at a cost of $1 quadrillion) is a society of 
pervasive surveillance where everything you say and do is public knowledge and 
searchable. This requires solutions to the language and vision problems. This 
will be practical once we have a million-fold decrease in the cost of 
computation, based on the cost of simulating a brain sized neural network. It 
could occur sooner if we discover more efficient
 solutions. So far we have not.

If you are selecting a language for self improving AI using a genetic 
algorithm, then you need to select a language near the boundary between 
stability and chaos, one with a Lyapunov exponent near 0. All programming 
languages (including LISP) have positive exponents (are chaotic). A small 
change to the source code (i.e. randomly flipping one bit) produces on average 
a large change in behavior that is rarely beneficial. A language is borderline 
stable if a random change is amplified by X with probability 1/X. Natural 
language and DNA both have this property. The ratio of helpful to harmful 
changes resulting from either random bit mutations or swapping large blocks of 
code is high enough to make the approach practical.

If you are selecting a language to implement language or vision, good choices 
are C, C++, and assembler. The primary concern is efficiency and the ability to 
make good use of underlying parallelism in the hardware. The choice will 
probably be less important as hardware gets faster.

-- Matt Mahoney, [EMAIL PROTECTED]




-------------------------------------------
agi
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=8660244&id_secret=117534816-b15a34
Powered by Listbox: http://www.listbox.com

Reply via email to