In an interview with Technology Review, the director of Intel Research, Andrew Chien, looks ahead to what happens when, in the not so distant future, we have “machines with tens or hundreds of [processor] cores perform[ing] trillions of operations every second.” What particularly excites him about what he calls terascale computing is “the ability for devices to understand the world around them and [infer] what their human owners care about.” He foresees pocket-sized “intelligent systems” that, using various sensors and tapping into powerful machine-learning algorithms, will continuously monitor our physical movements, analyze our speech, sense our moods, and anticipate our needs:
In order to figure out what you’re doing, the computing system needs to be reading data from sensor feeds, doing analysis, and computing all the time. This takes multiple processors running complex algorithms simultaneously. The machine-learning algorithms being used for inference are based on rich statistical analysis of how different sensor readings are correlated, and they tease out obscure connections. Right now these algorithms work on large systems built for a specific purpose, and it takes a PhD to get these things to work. We are looking forward to having these algorithms be in an API that you can call on, like a platform service which is as reliable to access as a file system. This way, the average programmer without a PhD can make use of these machine-learning algorithms.
At some point, one assumes, the programming of the device and the programming of the person would become indistinguishable.
So how long till this brave new world begins to emerge? Not long at all, says Chien: “Within five years, I think you’re going to see significant advances in performance. You’ll see demonstrations in the research world that are credible. I think the mainstream marketplace could pick up on it three years later, but at that point it’s hard to predict. The precursors for this technology are all there, though, and I see a huge need for it.”
What about glycemia? Isn’t that a key element of how wiches?
Que is mas macho, “annoyingly lifelike wafers of silicon” o “baseball bat?”
“Only an expert can deal with the problem,” -l.a.
-t
(Point being, I think that threshold is out there, not far off, already pointed at by things like iPod crime waves. Good riddance to bad rubbish lies ahead.)
Chien describes these “intelligent systems” as “understanding” what is going on around them and uses the word “infer” as though they possess consciousness and have the ability to reason. The only machine capable of this is: the human brain. It consists of hundreds of billions of core processors(neurons) working in massive parallel networks that are self organizing. The system his is describing consisting of a few hundred or thousand processes running programs analyzing sensor information falls very short of this. Its reminiscent of all the 1990s AI hype and the “intelligent web agents” of the dot com era – all of which crashed when Cyc, the symbolic “common sense” AI project, failed.
That’s not to say that systems like this don’t have great potential in applications like automobile and aircraft safety or childcare. IMHO, he’s being a little over optimistic and a bit intellectually dishonest to ascribe self awareness to this kind of thing. Machines capable of independent reasoning and consciousness will probably resemble the human brain in complexity and structure rather than a bunch of image analysis software running on a super computer in your pocket. And, will probably not be invented by anyone at Intel if this the best they can do! ;)
“The ability for devices to “understand” the world around them and [infer] what their human owners care about.”
The overall enthusiasm emanating from the statements made by Andrew Chien is embarrassingly naive! There will be some specific applications that might have limited uses but the above statement is just complete rubbish! It might help to clearly define what he means by understand!
I suspect that Mr. Chein is caught up in his own spin and shame on you for propagating what particularly excites him.
The battery is dying on my pocket-sized “intelligent system”, I . . . . . I am . . . . . . . . . . . . . . . . A .l . .a. . . n.
and shame on you for propagating what particularly excites him
I merely do what the machine commands.
NO!
It’s artificial intelligence trying to make a comeback. The biggest white elephant in the history of computing returns. Reaches for the vomit bucket.
Unfortunately, there’s a whole the herd out there. I call this kind of thing “intellectual crack” because pseudo–scientific claims like his have the same degenerative effect on scientific thinking that the drug crack has on the general population.
If the Intel R&D department’s repackaging of 20+ year old AI technology depresses you, check out physicist Peter Woit’s blog on String Theory called “Not Even Wrong”:
http://www.math.columbia.edu/~woit/wordpress/
He’s followed up and written a book about the time wasted by physicists on this intellectual pursuit that has produced practically nothing and wasted careers. Here’s a short review here the from Publishers Weekly Sept. 2007:
“String theory is the only game in town in physics departments these days. But echoing Lee Smolin’s forthcoming The Trouble with Physics (Reviews, July 24), Woit, a Ph.D. in theoretical physics and a lecturer in mathematics at Columbia, points out—again and again—that string theory, despite its two decades of dominance, is just a hunch aspiring to be a theory. It hasn’t predicted anything, as theories are required to do, and its practitioners have become so desperate, says Woit, that they’re willing to redefine what doing science means in order to justify their labors.”
At the advent of the 21 century, are AI and String Theory the pinnacle of human intellectual development? What’s next? Windows Vista with crop circles? It’s frightening to me that intelligent people are jumping down the rabbit hole after stuff like this especially since it appears the rabbit is carrying a crack pipe in his paw: “Oh, my ears and whiskers ..! do I have enough billable hours this week?!”
For anyone with an I.Q. above 100, I expect that one of the great flaws in the movie Terminator is when Reese says Defense network computers. New… powerful… hooked into everything, trusted to run it all. They say it got smart, a new order of intelligence.. Uh huh. Not just good enough to beat humans at chess. Not just a HAL 9000. Not even just Marvin. They just “got smart”. Uh huh. The flaws in the Terminator sequels are even worse (as most sequels are).
Perhaps you are thinking of the predicted technological singularity. What Ray Kurzweil talks about will, in a sense, happen, but it will not be because of single machines. It will be because of new global networks based on mutual trust between groups of humans, many existing examples of which already rule our lives (see below).
Nick, I already take you to task over in your later entry Neurotic bots rule. But let me make a concession: while computers of the future with neither rule man nor will they read our minds, they will “define the truth”. Yes, in the future, the Internet will define the truth. Especially for internet-obsessed couch potatoes who do not have their own scientific laboratories or network of first-hand witnesses to verify assertions of fact (which even now, for all practical purposes, is everybody). Just as our stable network of large banks defines how much money we have, just as the carefully-designed rules where a domain name will actually take us, just as the root certificates secure our internet connection that allows us to do banking online, so a hierarchy of Truth Servers will define the Truth in terms of the facts that are the safest for devotees of rationalism to believe. Instead of money or domain name registrations or the private keys of the root certificates, Truth Servers will deal in assertions of fact, perhaps with statistical ratings of some notion of “certainty”. Hopefully it will include enough checks-and-balances to avoid being taken over by some potential tyrant or benevolent dictator. Hopefully, it will be fault-tolerant, stable, and to some reasonable degree, tend to converge on internal consistency just like the well-maintained relational databases that store the records that rule our individuals lives with credit scores and the like (shades of the film Brazil). That network will still be managed and ruled by humans. One memorable example of a coup involving such a distributed network was a stunt pulled by Jon Postel several months before he passed away. I predict that that such a future network will often influence (to the point of dictate) what rational humans believe. That is not being ruled by machines: that is surrendering to a larger network ruled by humans. How does the guard in “Brazil” put it? Don’t fight it son. Confess quickly! If you hold out too long you could jeopardize your credit rating. Or does it already work that way? Anyway, you can take that to the bank.
Laws change over a period of time.
Worldwide Mind-reading system has been around for more than 30 years. What are we hiding ? Who are we hiding from ?
So is it time to de-classify Mind-Reading system ?
REMEMBER: THEY ARE YOUR TAX DOLLARS and THERE CAN BE NO ACCOUNTABILITY WITHOUT TRANSPARENCY.
Should Mind-Reading system be de-classified ?
THINK and vote at: http://www.geocities.com/mind_reading_system/