Monday, June 18, 2007

Machine intelligence

James D. Miller writes very perceptively on the rise of machine intelligence:
And, within 50 years machine intelligence will be so much vaster than human intelligence (unless we can somehow use Moore's Law to improve our brains) that it will dominate our economy and produce wealth beyond our current imagining. When this occurs we will achieve what has been described as a technological singularity.

The single greatest problem with the economics profession is its near total ignorance of the possibility of a technological singularity. To partially remedy this my soon to be published introductory microeconomics textbook concludes with a discussion of the singularity.
In his book The Singularity Is Near, Ray Kurzweil makes a similar argument about economics, in that the rise of artificial intelligence will create increases in productivity greater than the world has seen before.

In this paper, Robin Hanson writes of projected doubling times in world GDP growth rates of two weeks:
The suggestion is fantastic, namely of a transition to a doubling time of two weeks or less sometime within roughly the next century.

One might think this suggestion too fantastic to consider, were it not for the fact that similar predictions before previous transitions would have seemed similarly fantastic. We should also keep in mind that investors in U.S. stocks seem to be betting that a large change in the nature of the economy is likely soon. They may well be wrong, but this paper shows that the empirical case against such expectations is not as clear as one might expect from empirical analyzes looking at the last century or so. From a purely empirical point of view, very large changes are actually to be expected within the next century.
At the time of Hanson's paper, the internet bubble was in full swing, and its collapse has led many to pour scorn on the idea of fundamental changes to the economy. Yet these changes have kept chugging along, regardless of the stock market.

Huge changes in the economy are undoubtedly on the way. That is, if only we can keep the United Sates intact, something that the economics profession seems intent on preventing.

Labels:

5 Comments:

At 6/18/2007 11:15:00 AM, Blogger Tim R. Mortiss said...

There's a cultish, almost apocalyptic quality to the whole idea of the Singularity. "The Rapture of the Nerds", as one science fiction writer put it. What do serious artificial intelligence researchers say about it?

Assuming the Singularity actually arrives, there would be things more important than the economy to think about. Simply put, mankind would become obsolete.

In the best case, what would replace us could be understood as "better" or "more evolved" in some sense but, frankly, I can't relate to hyperevolved syntetic intelligences and I don't want my descendants to become ones.

In the worst case, the Singularity would pave the way to uninimaginable horrors.

 
At 6/18/2007 11:46:00 AM, Blogger Dennis Mangan said...

Tim: I think that some of your statement is based on a misunderstanding. Perhaps there is a "rapture of the nerds" quality to the idea, but Kurzweil bases his arguments on trend extrapolation only. Just as no one 100 years ago could have foreseen where technology and capitalism would lead us, it's hard to see where it will go from here. Kurzweil merely gives his best guess, and his evidence and reasons are compelling. And he does acknowledge possible downsides, like the grey goo scenario. The point is, nothing short of a worldwide clampdown on technology and capitalism can stop it. You won't have to "relate to hyper-evolved synthetic intelligence"; it will be part of you.

 
At 6/18/2007 12:37:00 PM, Blogger Tim R. Mortiss said...

"You won't have to 'relate to hyper-evolved synthetic intelligence' ; it will be part of you."

My reptilian brain is part of me too, but I doubt cocodriles can relate to me in any way except as a tasty morsel.

Imagine the following scenario: a protozoa has access to, and controls, a brain with human-level intelligence. And a cultivated one too, who reads Proust and writes poetry in iambic pentameter in his spare time. The protozoa only uses the brain to find the densest concentrations of nutrients. Do you consider this situation sustaniable? Don't you think the evolved brain might want to change the situation a little, perhaps disposing with the puny protozoa altogether? Excuse me if my examples seem far-fetched, but it was Kurtzweil who first made the leap into far out speculation.

Are the mental attributes which we regard as distinctively human (sense of self, empathy, and so on) inevitably tied to intelligence? Not neccesarily, in my opinion. Will intelligences with those attributes manage to thrive in a post-Singularity scenario? Maybe not.

By freeing intelligences from the contraints imposed by biological bodies and darwinian evolution, the Singularity would create an bewildering array of diverse mentalities. If Steve Sailer is right and diversity creates distrust, does that mean that the post-Singularity world would be mired in an eternal war between mutually unfathomable minds?

In fact, is Singularity evangelism compatible with any form of conservatism? Although, come to think of it, there are mormon transhumanists...

 
At 6/19/2007 01:33:00 AM, Blogger J. said...

No, humanity will not become obsolete. Please extrapolate into the rapture-like future the number of Mexican Grass Mowers in California.

But probably you dont mean THAT humanity.

I am being sarcastic, of course. This blog is so humorless!

 
At 6/20/2007 09:21:00 AM, Anonymous beavis said...

i saw a movie on A.I. once called Demon Seed or something.

at the end the evil super computer that tried to take over everything who used the voice of the Man From UNCLE guy was defeated.

they just disconnected his duracells, so im not worried.

 

Post a Comment

<< Home