[Carlhole]

US Air Force:

8/10/2009 - HANSCOM AIR FORCE BASE, Mass. (AFNS) — The convergence of “exponentially advancing technologies” will form a “super-intelligence” so formidable that it could avert war, according to one of the world’s leading futurists.

Dr. James Canton, CEO and chairman of the Institute for Global Futures, a San Francisco-based think tank, is author of the book “The Extreme Future” and an adviser to leading companies, the military and other government agencies. He is consistently listed among the world’s leading speakers and has presented to diverse audiences around the globe.

Canton seems familiar with the Singularity concept and views the US as rushing towards an unchallenged status:

“The superiority of convergent technologies will prevent war,” Doctor Canton said, claiming their power would present an overwhelming deterrent to potential adversaries. While saying that the U.S. will build these super systems faster and better than other nations, he acknowledged that a new arms race is already under way.

I'm reading Wired For War right now, a book about AI and robotics on the battlefield. The military accounts for about half of the total R&D for a variety of projects ranging from insect-like spy drones to Hummers that drive themselves to pilotless drones. Many of the projects are in a nascent stage, but they already have incredible capabilities. The drones have been very successful over Iraq and Afghanistan.

Eventually, The Air Force would even like to develop pilotless fighter aircraft so they don't have to worry about accommodating that blob of delicate jelly sitting up front anymore.

I'm glad to see that I have the Air Force is on the machine-intelligence side of the debate!

The US military could eventually create a superintelligence, but it will be a different beast than the billions of humans that came before it. A superintelligence is something fundamentally new. A superintelligence can hold bigger ideas in its brain than you can. Human working memory can only hold 5-7 items at once, a superintelligence’s working memory might be able to hold millions of complex symbols simultaneously. It would be like comparing a man with a shovel to one of today's earth-movers.

Who knows what the Air Force or DARPA has going on under its blacked-out projects?

[Kristen]

Although I like your optimism, you need to be more specific as to how much a human can remember. 5-7 items for about 8 seconds, without rehearsal.

Also this would we even be able to understand such super intelligence?

[dinopello]

A superintelligence can hold bigger ideas in its brain than you can. Human working memory can only hold 5-7 items at once, a superintelligence’s working memory might be able to hold millions of complex symbols simultaneously.

The Magic Number 7 plus or minus 2

I'm skeptical that this characteristic will be overcome in any significant way. I believe it is somewhat related to a fundamental characteristic of complexity and to the ability of learning (forming models) by induction. The decomposition of complex problems, or chunking is required because enough evidence is rarely available to support the formation of valid models of high dimensionality, so humans form these lower dimensionality groupings which are then combined. Many of the successful AI techniques mimic this type of composition.

This paper by the Barron's, although 20 years old now, is a pretty good overview of empirical techniques in use for machine-based learning (PDF).

[Carlhole]

A superintelligence can hold bigger ideas in its brain than you can. Human working memory can only hold 5-7 items at once, a superintelligence’s working memory might be able to hold millions of complex symbols simultaneously.

I'm skeptical that this characteristic will be overcome in any significant way.

Well, the Air Force isn't skeptical. The military is ALL OVER this stuff in a very big way. Who knows what black projects they have going on.
I find The Blue Brain Project (non-military) to be the most fascinating one so far because it is developing a model which other researchers can continually build upon. It will incorporate EVERYTHING that is known about neuroscience to date. The computational requirements to do this study are absolutely mind-blowing.

I saw another article today about it.
Human Brain Could Be Replicated In 10 Years, Researcher Predicts

Today's technology is already highly sophisticated and it allows us to reverse engineer the brain rapidly." An example of the capability already in place is that today's robots can do screenings and mappings tens of thousands of times faster than human scientists and technicians.

Another hurdle on the path to a model human brain is that 100 years of neuroscience discovery has led to millions of fragments of data and knowledge that have never been brought together and exploited fully. "Actually no one even knows what we already understand about the brain," says Professor Markram. "A model would serve to bring this all together and then allow anyone to test whatever theory you want about the brain.

Advanced AI is going to get away from the limitations of serial-processed software and move towards vastly parallel systems which are copied from the information processing designs that Nature has already produced.

No one can convince me that it is impossible for scientists to study and learn about the biology of neurons and the information-processing patterns of various life forms. I mean, if thousands of inexpensive drones flying over Afghanistan and Iraq were as independent as even a dragonfly, that would be a butt-kicking weapon.

There are a bunch of these projects going on all over the world now. Here's another:
Building a Brain on a Silicon Chip

FACETS has been tapping into the same databases. "But rather than simulating neurons," says Karlheinz, "we are building them." Using a standard eight-inch silicon wafer, the researchers recreate the neurons and synapses as circuits of transistors and capacitors, designed to produce the same sort of electrical activity as their biological counterparts.

[nobodypanic]

yay, skynet here we come!

[Arthur75]

This type of declarations is nothing more than a new version of the hope in the venue of the messiah.
AI is junk (or more precisely there is nothing special about some software qualified as AI and some other that isn't)
Most AI "experts" are boring idiots

[smallpoxgirl]

yay, skynet here we come!

I was gonna say. I saw a movie about this.

[lowem]

SkyNet is one possibility.

The other they are alluding to is a vast planetary- or galactic- scale hyper-intelligence that, sorry to say, is even further off than SkyNet is.

That's right folks.
It's easier to build a SkyNet than a Type 3 Civilization.

See, I can talk Singularity too

[dinopello]

Most AI "experts" are boring idots

Yea, those morans !

One difference of intelligent/learning systems is that the response of the system is not deterministic. If a robot put in the exact same situation makes a different decision because it's code was self-adapted during some intervening experience, then it does behave quite a bit differently than a conventional SW program.

[evilgenius]

So what happens once the capitalists angle for control realizing that they can exploit the rest of the world with impunity. Who is going to stop them? What, some form of democratic tribal faux-war game shake? How does a solitary man in the right appeal against such a super-intelligence? In what category do we put all of the charlatans and other complainers that are nothing but deluded self-important blips, but are still human beings, presumably with rights?

I have every confidence that we can develop AI. My fear is that we will use it to continue being who we are now, rather than to help us discover who we can be. The first reform is to reform the money structure, then we can mine the wonders of AI.

[Ferretlover]

The convergence of “exponentially advancing technologies” will form a “super-intelligence” so formidable that it could avert war, according to one of the world’s leading futurists.

HHHmmm... Don't think so. That would kinda sorta defeat the purpose for having an air force or any other military organization don't cha think? "Ok, all you countries sit up straight and raise your hands: How many of you are willing to dismantle all your militarys in favor of just having a super-intelligent technology?" (Good time to take a picture since there is no movement).
"Thought so."

[Arthur75]

Yea, those morans ! One difference of intelligent/learning systems is that the response of the system is not deterministic. If a robot put in the exact same situation makes a different decision because it's code was self-adapted during some intervening experience, then it does behave quite a bit differently than a conventional SW program.

No it's not, and the problem of technology currently is not related to that at all anyway

[smallpoxgirl]

I'm not sure how intelligence is going to avert wars anyway. Wars aren't caused by a lack of intelligence. They're caused by selfishness overwhelming our desire to behave intelligently. What's this thing going to become some sort of international marriage counselor? "Mr. Ahmadinejad, why don't you tell Mr. Obama how that made you feel."

[dinopello]

No it's not, and the problem of technology currently is not related to that at all anyway

What's not ? The thing about problems with technology is we'll concoct some new technology to fix whatever it is wrong currrently. The primary cause of problems is solutions. Or, put another way, a key benefit of solutions is that they create more problems to solve. Wait, that does not compute ...

[Cloud9]

Will that thing fold my socks?

[Carlhole]

See photo in Gasmon's post

Now, I find that deeply, deeply offending of our proud, state-of-the-art Air Force whose true operational capabilities I cannot even guess. Why do you hate the USAF?

[rangerone314]

See photo in Gasmon's post

Now, I find that deeply, deeply offending of our proud, state-of-the-art Air Force whose true operational capabilities I cannot even guess. Why do you hate the USAF?

Old WW2 saying

When the Germans are flying, the English duck.
When the English are flying the Germans duck
When the Yanks are flying, EVERYBODY ducks !!

Gasmon

Hey, stop picking on them... my dad was in the Air Farce!

[Carlhole]

Scientific American

Blue Brain isn't the only effort to reverse engineer the brain. The Defense Advanced Research Projects Agency (DARPA), the U.S. Defense Department's research arm, last year gave the IBM Almaden Research Center $4.9 million for a project called "SyNAPSE," an attempt to reverse-engineer the brain's computational abilities to better understand its ability to sense, perceive, act, interact and understand different stimuli. Although the brain is still not well understood, "there is enough quantitative data for us to be able to begin putting together the pieces," Dharmendra Modha, IBM Almaden's director of cognitive computing, said earlier this year at an event celebrating the Institute of Electrical and Electronics Engineers' (IEEE) 125th anniversary. Modha predicted that by 2018 computers will be able to simulate the workings of the human brain.

In Scientific American's 2008 Special Report on Robots, Ray Kurzweil, CEO of Kurzweil Technologies, Inc., proposed that the fastest way to reverse engineer the brain may be to study the real thing. One condemned killer has already allowed his brain and body to be scanned, Kurzweil points out, and all 15 billion bytes of the now-digitized inmate can be accessed on the National Library of Medicine's Web site.