[Rod_Cloutier]

With the upcoming release this week of new computer processors with 10 cores, I have to wonder if the emergence of artificial intellegence isn't too far away:
CPU

I would think something like Asimov's benevolent robots will help us sort out our social problems, and solve problems caused by human limitation. I saw an article today that people who 'believe in the coming of the singularity' are already standing in line to have computer chips embedded in their brains as soon as the technology is ready.

Nest big future

I must say all this talk of the comming singularity makes an old timer like me a little nervours about the emergence of a 'Borg' or 'Skynet' type of nightmare situation in the future. Especially since many young people claim the singularity will occur within the next ten years. This still within a period before we can expect the consequences of population overshoot and resource depletion to destroy modern society.

[Sys1]

No.

[Cloud9]

Not to worry, if the coming singularity is no more accurate than my Tom Tom it will be running around in a virtual cul-de-sac in Atlanta until its batteries run flat.

I for one don’t need or want to be saved from myself.

[Narz]

I think AI technology will help save us from ourselves. Our biological evolution alone simply isn't going to cut it. Humans simply aren't hardwired to think globally.

[Cloud9]

Ah I get it. It has been such a short time since we came out of the tall grass that we have not evolved to the point that we think about our individual actions and the impact those actions have on the environment. Inasmuch as we are all egocentric we as a species cannot grasp the big picture.

A machine on the other hand can process trillions of bits of data and can arrive at the best possible global solution, a solution that would benefit all species not just one.

I think that scenario has already been played out.

YouTube

[Narz]

Cause life always ends up just life in the movies, right? If all we can envision what has already been given us by the media we're truly in dire straights.

[Outcast_Searcher]

I've read all of Ray Kurzweil's stuff (at least his books), pounding the table that it is inevitable that we'll have human++ level A/I any time now -- just as soon as the silicon gets fast enough.

I've also read a lot of ivory tower philosophy that (rather pedantically, it seems to me) states that is impossible due to some logical point that is either silly, or that I'm too stupid or poorly educated to fathom.

So guess what -- WE WON'T KNOW IF "SMART" A/I IS POSSIBLE UNTIL WE CAN ACTUALLY BUILD THE DAMN THING AND SEE IF IT IS, FOR EXAMPLE, SELF-AWARE, can laugh at jokes, can make reasonable judgements, etc.

That may take another 20 years or more. Show me something with the intelligence, say, of a house cat, or better yet a crow (crows are very smart -- see Nova's "Murder of Crows") and I'll at least start to believe this is likely.

Until then, this is all like monks arguing about how many angels can dance on the head of a pin. You need some angels before you can draw any meaningful conclusions.

(OTOH, special-purpose A/I looks pretty impressive. Self-navigating cars have made amazing progress. Chess computers have FAR surpassed humans now. I have hopes that in 20 or 25 years that my car might be able to drive me around the city, and that in 30 to 35 years, a robot might be able to be my nurse, should I need one).

[AgentR11]

So guess what -- WE WON'T KNOW IF "SMART" A/I IS POSSIBLE UNTIL WE CAN ACTUALLY BUILD THE DAMN THING AND SEE IF IT IS, FOR EXAMPLE, SELF-AWARE, can laugh at jokes, can make reasonable judgements, etc.

But testing self-awareness in a machine, or even another species of animal will be, and is, very subjective based upon the perceptions of the evaluator. Does code that evaluates spoken text against a database of 5 million punch lines, indicate an understanding of humor? Does failure to laugh at a joke that *I* think is funny, imply lack of understanding of humor?

I think yall have the A/I thing upside down. Human brains are cheap; people across the world put their brains in the service of others for a pittance. So recreating a machine version of that is not only difficult, but a guaranteed money loser, useful only in the sense of "lets see if we can do it once..." OTOH, code and processors able to navigate hazards and perform tasks in environments unsuited to flesh and blood; definite win. Imagine how much easier it would have been to work on that oil spill if right at the start you toss a dozen semi-autonomous AIBots over the side, and they set about the task of taking pictures, clearing pipe, and later manipulating large objects and tools. Or the current problem in Japan with shielded versions of AIbots able to drag hoses around, connect cables and pipes, etc; and if the radiation kills it, its chassis remains usable, just rip out the circuits and install new ones. Modest iterations of this type of AI are already in play on the martian rover Opportunity. Machines can be programmed to make relevant decisions concerning their operation, without the need to address the issue of intelligence or awareness.

The other side of this puzzle is remote commanding, where human input in a safe location causes motion and actions in a remote location... predator drones anyone? "He who blows up the fewest Taliban today buys the coffee!" This has the advantage in that the responsible party remains clear; where with an independent AI, its quite vague.

[AgentR11]

nb... the future is not in making faster and faster processors. Its making more and more independent processor cores able to function on the same buss.

Not unlike the human brain, a bajillion really gimpy processors hooked up on the same buss creating an IO/signal processing machine of unmatched power; all for the low cost of coffee and a poptart.

[ian807]

Some of us, perhaps. AI is a tool. If we're smart enough to use it to survive, we'll do OK. Some humans will use it for hostile purposes, at first, until there's no reason to do so. Others will hybridize with it and they will speciate. What they will do at that point is unpredictable. The motivations of a human/hybrid AI will be too alien for us to understand.

Remember AI, of itself, has no motivation. It's not organic. It doesn't care if it survives or not, or whether you do. Any assumptions you make about it that makes it human-like, even mentally, are almost certainly wrong.

[Rune]

EE Times: Vicarious AI Passes Completely Automated Public Turing Test

Vicarious who?

Vicarious is a three-year old California Flexible Purpose Corporation (thus the FPC after its name), which instead of maximizing shareholder value, like a normal corporation, is aiming to fulfill a singular purpose -- to solve the important algorithmic problems behind building a human-like AI. For Captcha, its already succeeded, but Vicarious's long-term goal is to generalize its Captcha AI into a complete robotic brain that is as smart as a human in all areas of sensory perception.

"Our goal is to combine insights from neuroscience with modern machine learning techniques and cast them into mathematical algorithms that are just as intelligent as humans," Phoenix told us.

The three-year-old startup was founded by Phoenix, formerly entrepreneur in residence at Founders Fund, and co-founder Dileep George, formerly chief technology officer of Numenta. Vicarious is running on its second round of funding. The first seed round in 2010 was $1.1 million, and its Series A for $15 million was just completed last year, giving the six-person company the time and money it thinks it needs to fulfill its mission of casting human intelligence into mathematical algorithms, which it expects to achieve in five to seven years.

How it works

Vicarious's secret weapon is what it calls a recursive cortical network (RCN) -- a machine learning framework that embodies the structural and computational power of the brain's neocortex.

"Our algorithms express what we think is going on inside the neocortex of the brain," says Phoenix.

Thus the big difference between Vicarious's models of the brain and the neural network models of the colossal efforts of the European Union's Blue Brain Project and the US Defense Advance Research Project Agency's (DARPA's) Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE), is that Vicarious is not trying to mimic all the details of the biological brain, but only its overall functionality.

"We are not trying to emulate the physical brain -- like the EU's Blue Brain or DARPA's SyNAPSE program," said George. "Rather we are attempting to understand what the brain is doing and encoding that into mathematical algorithms. This recursive scaffold allows our algorithms to understand how objects in the world are related to each other and what they are used for, and that allows us to skip over all the complexity that is built into models trying to emulate all the details of the brain's connectivity."

I'm tellin' ya man... they're coming... to a competitive intellectual sphere near you.

Frankly, I am utterly amazed at the sheer stupidity of most people. Whilst, simultaneously, utterly stunned by the brilliance of others.

I have to wonder what part of the human mind is least able to be duplicated by machine intelligence. I don't think that there is ANY part of it that is not replicable by some digital analog.

Probably, the parts of the human mind most difficult for machines will be the talents that ANY dumbass can do - like see things and identify them. Like how a human mind can instantly see an apple in all the millions of different ways you can see one.

Conceptualizing a multi-dimensional universe and its quantum mechanical workings might be a whole helluva lot easier for the machines.

[copious.abundance]

I have this personal theory (based solely on a hunch, I admit), that machines (as we know them now) can become as intelligent as humans (or nearly so, and maybe more intelligent in some aspects), but they will not become conscious until we make fully intelligent analog computers.

My reasoning is thus: To become conscious, one needs to be aware of one's self - correct?

And what is that self but a physical object?

Currently, computer circuits (as I understand them to work, correct me if I'm wrong) are basically inanimate silicon objects with pulses of electricity passing through channels. Silicon chips could never become conscious because the electricity the chips manipulate is basically disembodied electricity. Disembodied electricity is never going to become self-aware.

In a brain, on the other hand, the physical substrate is itself charged with electric pulses. Thus, the physical object and the electricity are "one." The brain becomes aware because the physical substrate is itself "electrified."

So, in order to produce a conscious computer/robot, according to my theory, we will have to make analog computers. A digital computer could become intelligent, but it will be a non-conscious intelligence.

However, I admit I don't know enough about how circuits and chips work to have much confidence in my little theory. I definitely could be wrong about this.

[Rune]

Most certainly everyone will be wrong about it in one way or another because eventually machine intelligence will create its own evolutionary growth path - at first, mostly coming from human inputs but, increasingly, taking on design jobs for the NEXT generation of intelligent or near-intelligent machines.

and no one can tell where this evolution will lead.

[AgentR11]

I think you'd have a hard time distinguishing true consciousness from simulated consciousness; but simulated consciousness might be more than sufficient. I suspect Rune has it right, that the hard things may be the things that are trivial for even the institutionalized 70IQ... What *IS* an apple. Why do I know "apple" when Iook at a red delicious, or granny smith, or even its more distant relations, but then see a pear, and know, "NOT apple". Even if they're disfigured or damaged, I still know "apple" and I know "pear".

The machine could easily know every human language word for "apple"; and still fail at knowing one apple vs one pear.

I do wonder a bit if its really a worthwhile pursuit though; humans are cheap, 7 billion of us, many willing to work themselves to the bone for a couple bucks a day. It could be that the valuable contribution turns out to be the interface between the digital machine and the biological, signal processing supercomputer that is the human brain (that runs on a few hundred grams of sugar and flour a day..)

[Rune]

We saw a machine defeat the worldLs chess champion, then we saw a machine defeat the top Jeopardy! players, then we see that Google cars are much safer than human drivers, then it will be some other skill, then another one... we will continue to see these things develop.

It will go somewhere.

I could imagine a machine that does only electrical circuit design and does very, very good work. Or a machine that does accounting very, very well be it personal or corporate - machines that learn a vast quantity of data, essentially.and, of course, machines that analyze vast quantities of data tha no human mind possibly could.

[Keith_McClary]

Scientific American

New brain-mimicking artificial intelligence software can read security images called CAPTCHAS as well as people do, its developer says--but scientists are skeptical
By Susan Kuchinskas
Luis von Ahn has heard it all before. As co-inventor of the CAPTCHA, those annoying images composed of wiggly letters and numbers that Web sites use to make sure you’re a human rather than a machine, von Ahn has received as many as 50 claims over the past decade of ways to beat his program.

Make that 51.
...
Vicarious's CAPTCHA-solving demonstration is an example of "narrow artificial intelligence," a technology that can match or even exceed human performance on a narrowly defined task. IBM's chess-playing Deep Blue is another such example. But Vicarious insists its computer perception software is the foundation of an AI that will learn the way humans do—by experiencing the world around it, principally via vision, and then identifying patterns. "If an algorithm solves vision in general, it is not narrow AI, it's a general AI system,” says, Dileep George, also a Vicarious co-founder. “We are working on a general algorithm for solving [the] vision problem, and CAPTCHA is a stepping stone to that.”
...
CAPTCHA inventor von Ahn, an associate professor of computer science at Carnegie Mellon University, doesn’t seem excessively worried. It is hard to determine exactly how much better Vicarious's technology is than other work in the field, he says. The Vicarious approach, which relies on visual perception, is in line with current thinking about AI, according to von Ahn. "Many artificial intelligence researchers spend most of their time dealing with perception,” he says. “It's believed that our own intelligence derives from our visual cortex.”

Even if it proves to be a technological dead end, "the one nice thing about the approach of using computer vision is that at the very least, it has applications," von Ahn notes. For example, technology based on Vicarious’s system might someday give a self-driving car the ability to identify pedestrians straying onto a roadway.

Those darn skeptical scientists can't see the brilliance.

[SeaGypsy]

Even if it proves to be a technological dead end, "the one nice thing about the approach of using computer vision is that at the very least, it has applications," von Ahn notes. For example, technology based on Vicarious’s system might someday give a self-driving car the ability to identify pedestrians straying onto a roadway.

Those darn skeptical scientists can't see the brilliance.[/quote]

AI is well outside my field of expertise but I don't think 'someday' applies to the above concept- the technology to avoid pedestrian collisions, like collisions in general- already exists. Someone is already selling a car with 'city mode braking' automatic collision avoidance as a standard feature in an entry level car. The way these things are going they will be cheap and mandatory on all new vehicles soon.

One of the best things I can think of with this stuff is being able to lighten the commuter fleet very significantly without compromising on safety. Collision avoidance, new tech very high ratio output motors, some cool stuff. I guess AI could make a huge difference to traffic management also. Interesting times and not just in the Chinese curse sense

[dorlomin]

We saw a machine defeat the worldLs chess champion,

Number crunching.

We do not have a single machine that can come anywhere near a cockroachs ability to deal with its environment. We dont even have a machine that deal with the balance like a cockroach.

A computer is no closer to sentience than a 2 stroke lawnmower.

[SeaGypsy]

A 2 stroke lawnmower with a small degree of sentience could mow your lawn better, safer, more efficiently for perhaps very little extra cost once these components become commonplace. I'm not a 'singularist'- I can never see the time where 'I' actually exists in digital form as such. However I don't think this matters much, nor does supposedly super mega machine intelligence per-se. Machine intelligence has some acutely interesting applications. Nothing wrong with considering them IMO.

[vision-master]

I have to wonder what part of the human mind is least able to be duplicated by machine intelligence. I don't think that there is ANY part of it that is not replicable by some digital analog.

Emotions for one, spirit for another.........