On Wed, Oct 15, 2014 at 1:04 AM, Piaget Modeler via AGI <[email protected]> 
wrote:
>
> So the internet is an AGI, is that your claim?
>
> Hmmm.... that's dubious.

What I find dubious is the idea that all we need to do is find the
right algorithm for modeling human learning and then train it with
human knowledge, and all our problems will be solved. Ben Goertzel has
been working on this problem since 1997, first as Webmind, then
Novamente, and now OpenCog. Doug Lenat thought it was just a matter of
encoding common sense in structured form, and has been working on Cyc
since 1984. Marvin Minsky and Ray Kurzweil spent decades on this
problem starting in the 1960's, but have since scaled back their
ambitions. Alan Turing described the approach in 1950, estimating it
would be solved by 2000, requiring 10^9 bits of storage but without
any need to increase processing speed. Yet here we are in 2014, still
paying people $70 trillion per year to do work that machines aren't
smart enough to do.

It's not that we aren't making progress. In fact, world GDP is a lot
higher than it would be if our economy weren't so dependent on the
internet. Many jobs wouldn't even exist otherwise. We don't think of
the internet as AGI because it doesn't pretend to be human. Never mind
that it can answer far more questions than any human could, and a lot
faster. The internet isn't intelligent as long as we take credit for
creating and using it. Google isn't smart. It just makes everyone
smarter.

Given the enormous incentives to automate human labor, I'm not sure
why we even want AGI. A lot of the development work in passing the
Turing test goes into slowing down response time and inserting errors
to make chatbots look more human (a problem Turing was aware of). Why
would we even want to build practical machines with human weaknesses,
like poor memory and arithmetic skills, emotions, or a need to take
time off to sleep?

So you're right. The internet is not AGI, even as it surpasses humans
in knowledge, computing power, and every skill that is important to
the economy. But I would keep an eye on it, just in case you were
expecting AGI to arise somewhere else.

-- 
-- Matt Mahoney, [email protected]


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