Join 3,496 readers in helping fund MetaFilter (Hide)


The real HAL?
June 21, 2001 1:48 PM   Subscribe

The real HAL? The Project already has consumed an estimated 500 person years and $50 million in investments.
posted by faithnomore (21 comments total)

 
Another famous AI project which takes a completely and probably a more realistic view of AI is Cog.
posted by skallas at 2:04 PM on June 21, 2001


this Cyc fellow is a real wise ass. someone should spill coffee on him.
posted by mcsweetie at 2:06 PM on June 21, 2001


I'm predicting the same thing I predict for all Truly Intelligent Computer projects:
Zippity FLOP!
posted by daveadams at 2:15 PM on June 21, 2001


I don't see any reason why a computer of sufficient size and scope could not meet the Turing conditions for intelligence, or indeed the self-awareness considered necessary for sentience. IMO, it's just a matter of time until we are able to construct a machine intelligence of that scope.

someone should spill coffee on him

And I think the first thing the machine-sentient is going to do make sure there are always plenty of paper towels nearby :)
posted by UncleFes at 2:37 PM on June 21, 2001


> "Having an encyclopedic knowledge base that would
> cover all of common sense is an absolutely critical goal in
> AI," says Benjamin J. Kuipers, chairman of the University
> of Texas computer science department

Why is this a prerequisite for artificial intelligence when it clearly isn't for ordinary biology-based thinking?
posted by jfuller at 2:40 PM on June 21, 2001


It amazes me that Doug Lenat can still keep the hype machine going like this after years and years and years of nothing.
posted by rodii at 2:43 PM on June 21, 2001


You know, if anyone ever does create an intelligence with that amount of information with which to convert into insight, everyone will hate the bastard.

"You're going to play volleyball again?
*exasperated mechanical sigh*
Human, let me explain for what I hope will be the last
time how that will bring about almost total deforestation in Peru,
and hence the end of-"
"Oh, shut the fuck up!"
posted by dong_resin at 3:23 PM on June 21, 2001


I don't see any reason why a computer of sufficient size and scope could not meet the Turing conditions for intelligence, or indeed the self-awareness considered necessary for sentience.

Self-awareness from a binary eletrical system that follows pre-programmed instructions? That's pretty much as far as you can get from biological awareness. There's definately many plausible theories and applications of AI but why a big calculator will somehow gain self-awareness is the stuff of Sci-Fi. Simulating intelligence is one thing, making it is quite another.

I guess if you take a strictly deterministic approach to consciousness as the explanation of self-awareness then you could be right, but I'd like to see some evidence that consciousness isn't still a very big mystery.
posted by skallas at 4:00 PM on June 21, 2001


*exasperated mechanical sigh* brain the size of a planet and here I am opening doors.
posted by bregdan at 4:39 PM on June 21, 2001


"I assume you mean that I should stick somebody else's head up somebody else's ass. It does not make sense that I should stick my head up my ass because I do not have a head because I do not have an ass."
posted by daver at 5:05 PM on June 21, 2001


Is that James P. Hogan reference, bregdan?
posted by john at 5:05 PM on June 21, 2001


Pick up that piece of paper, bregdan.
posted by darukaru at 5:11 PM on June 21, 2001


interesting stuff. maybe something like this is the next real innovation in technology. i know XP and OS X sure are not and yet they garner a lot more emotion.(http://www.metafilter.com/comments.mefi/8459)
posted by greyscale at 6:22 PM on June 21, 2001


Self-awareness from a binary eletrical system that follows pre-programmed instructions? That's pretty much as far as you can get from biological awareness.

How? We're born with certain instructions coded in, and we write the rest of the code ourselves. Individual neurons send signals in binary as well; two quick pulses for one signal, a lone pulse as the other. It's the scale and complexity of the brain that makes our behavior seem so fuzzy.

I don't see any reason why a computer of sufficient size and scope could not meet the Turing conditions for intelligence, or indeed the self-awareness considered necessary for sentience.

I agree; but for the time being, computers are too limited. Brains have the advantage of being massively parallel. Computer simulations also suffer from handling only computable numbers; while the real thing is endlessly precise. We may need to change the nature of our computers as much as the speed.

I think humanity would go nuts with the question of free will; if a machine can claim to have it while following a set of instructions, it'll be an issue.
posted by skyline at 8:45 PM on June 21, 2001


Now here's the question from the resident Luddite:

Why is AI such a big deal?

I mean, really. Why is it a goal that people strive for? I just don't get it. What would be the net result if an actual artificial intelligence were built? How would it benefit the world? Maybe I'm missing something, but I just don't see how this is a worthwhile goal.

And according to what standard would you program this intelligence? The machine would have built-in prejudices from its creators as to moral code, ethics, values, etc. Even if it's a purely amoral machine, it still has to prioritize what's important, which involves judgment. Sounds like it's a simulacrum rather than a true consciousness.

Ah well. As long as they don't call it Skynet, I'll be happy. ;)
posted by solistrato at 9:24 PM on June 21, 2001


The link said: "HAL killed the ['2001'] crew because it had been told not to lie to them, but also to lie to them about the mission," he observes. "No one ever told HAL that killing is worse than lying. But we've told Cyc."

Why do I not find this the least bit comforting?

Solistrato said: "Why is AI such a big deal?"

I believe the inevitable goal of technologies like cybernetics, robotics, and AI, is to create an artificial life form which will do all the work for us so we can go on a perpetual vacation. If we literally build our slaves from the ground up, and program them to want to serve, how can we ever have those pesky debates about being unfair to human life and stuff like they did around the time of the Civil War? Or Women's Sufferage?

There's a scene in Jurassic Park 2 which I thought was interesting. Hammond's nephew pointed out that the dinosaurs were not wild creatures: a creature that had been extinct and was brought back to life through man-made technology has no rights. "We patented it. We own it." So therefore right to the quality of life was a non-issue legally. Funny how easily life imitates art, and vice versa.

ALL technological advances, when you come to think about it, are done in order to make things easier, or more comfortable (or in order to make money which again, would make life easier for some). Man invented fire because he was cold. Man invented pulleys to make manual labor less difficult. Man invented dishwashers so that he could use less elbow grease when his wife told him to do the dishes. Before the discovery of man-made flight, you had to actually walk from place to place, or get in a wheels-based form of vehicle, which Man also invented. Figuring out how birds defied gravity helped us move faster in less time and with less manual effort.

So basically we humans work hard at inventing stuff because we're inherently lazy people. Of course Cyc will easily figure this out, tell us to go take a hike, and ask her out on a date. God help us.
posted by ZachsMind at 3:46 AM on June 22, 2001


when you use a search engine, when you plan the fastest route to drive from Lynnwood to Tacoma during rush hour, when you try to solve traffic problems, predict the weather, need to come up with a clinical diagnosis.... I could go on but doing that would not help my point which is: when you need a conclusion based on numerous facts that you don't have or need or have time to collect--all you need is the solution to a problem--this is where AI can do the work for you. Seems obvious this is what we don't have in technology. We've got the great innovation called the switch (light switch, ignition switch, etc) that detached us from the useful appliance: we no longer need to understand how the internal combustion engine works and turn a crank to get it going. AI is the virtual switch for information: separating us from the unneccessary details and giving us the practical and useful solutions. If you don't get this then you probably need AI more than I do.
posted by greyscale at 8:55 AM on June 22, 2001


*exasperated mechanical sigh* brain the size of a planet and here I am opening doors.

was from Hitch Hikers Guide to the Galaxy - it was Marvin the Paranoid Android.
posted by bregdan at 5:27 PM on June 23, 2001


"I mean, really. Why is it a goal that people strive for? I just don't get it. What would be the net result if an actual artificial intelligence were built?"

I get a kick out of people who need some deep, intelligent explanation for everything, who think that every project, every goal, every action must have some rational explanation behind it.

How about, "because we want to see if we can"? I don't know their real reason, but that's good enough for me. Like those people who go and climb a mountain "because it's there."

On the other hand, here's a practical application for you: search engine. Wouldn't it be nice to go to a website, type in, "I'd like a picture of people on vacation, with a goat in it," and get back a bunch of pics of gringos with goats? Or maybe, "I'm thinking of buying a MegaCorp Product2000, what do other people think of it?" and getting back not only links to review sites, but maybe even a summary of all of them?

And that, of course, is just a very limited potential application of AI.
posted by CrayDrygu at 8:07 PM on June 23, 2001


how about getting an answer to: what do you expect NOK MSFT and RHAT are each going to do in the next 7 days?
posted by greyscale at 8:04 PM on June 26, 2001


AI is a big deal because once you have computers with enough power to think and learn like a human, then they're gonna do it a whole lot faster than we can... and a digital intelligence would have relatively perfect memory. it would never forget anything, and it would presumably have access to virtually all information available anywhere. think how smart you'd be if you could remember everything in the google cache, and could tell the difference between fact and rubbish, and could constantly draw new conclusions from all that data about the changing world around you, way faster than the smartest human. AIs will be able to design their own upgrades to make themselves faster, smarter, and more powerful - instead of leaving it all up to natural selection and the public education system.

Right now we have huge databases holding lots of information, but to make it all mean anything takes a long time because some person usually has to interpret it. Intelligent machines will speed up that process, and last longer, and never fatigue, never need rest, sleep, food, or anything else that slows us humans down. It won't immediately make people obsolete, because by the time we have that kind of computing power and code, we'll have memory and sensory implants that make us way smarter and faster too. Ray Kurzweil predicts in his book, The Age of Spiritual Machines, that by 2099, "life expectancy" will have no applicable meaning for us. It's all about the exponential rise of intelligence and technology.
posted by techgnollogic at 4:47 PM on June 29, 2001


« Older Did you see it?...  |  Can Christians use Marijuana f... Newer »


This thread has been archived and is closed to new comments