There's a POV on intelligence which is similar to the one regarding  the
"soul", "consciousness", "free will" - these are social "rewards", items
used for social discrimination, the ones who got it are "higher" and that's
their "objective" and "ethical" reason to have certain rights.

The "defintion" of intelligence is arbitrary only in cases, I am personally
fully aware what I want my thinking machine to be able to do, and what it
doesn't do yet/is not implemented  (or what the others' machines don't).

I will also remind, that "human level" and "humanity level" are confused. A
human-level AGI might be actually a super-human and "humanity-level" by the
beginning due to the humans input-output bottleneck which the machines
lack. The bar for the average humans is pretty low in most domains, also.


Matt>Right. We expect something special to happen when computers >surpass
human intelligence. As if such a thing had any meaning

Yes, it will be very spectacular and very quick...

For example, one computer alone will be able to code an operating system
that's functionally equivalent to a modern OS developed for decades, well
for the history of electronic computers, by thousands or millions of people.

It will do like this: flip-flop. The device drivers? It may design it in a
way that's compatible with the drivers of some particular OS and pick those
drivers.

Anyone here familiar with the OS? Complex?

I don't think so - it's complex and hard only for humans, and only at first
sight - it's big as number lines of code, otherwise it's just a literal
function of the CPU's architecture that's unfolded and multiplied for
multitasking, with some arbitrary decisions, some of which are a function
of human memory peculiarities regarding their working memory capacity and
how the functions are divided etc.

A thinking machine could experiment anyhow regarding the arbitrary
decisions. Given it could use virtual machines to test them, and that it
actually wouldn't need to do tests the way humans do, it could read the
literature on OS and pick the best practices.

It would write a novel or generate movies on the fly - humans also can do
it, everything is done on the fly, but the problem with humans is that we
are too slow.

All the software that has been created, the functional part of it, would be
possible to be generated in almost no time from scratch, any needed
software could be literally "spawned" or re-configured or ported
 immediately on demand by any computer with VLSI/AGI software on it.

And that's just one spectacular thing that will happen quickly after the
Versatile Limitless Self Improvers are once integrated and run.

I think it should come pretty soon, even if you just notice how much absurd
and inadequate for current technologies is the process of hand typing code,
using keyboards which are essentially the same like 19-th century
typewriters...  :-D

It is so inefficient and slow, even with predictive tools and interactive
help - a developer may be seeing an entire complex software system, but he
may need to type in and debug 30-50-100K lines of code, which may take an
age if done the silly conventional way.

The humans' slow fingers are the bottlenecks of Progress... ;)


*-- *
*=== Todor "Tosh" Arnaudov ===*

.... Twenkid Research:  http://research.twenkid.com

.... Author of the world's first University courses in AGI  (2010, 2011):
http://artificial-mind.blogspot.com/2010/04/universal-artificial-intelligence.html

.... Todor Arnaudov's Researches Blog: http://artificial-mind.blogspot.com



-------------------------------------------
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

Reply via email to