Do you think our technology can support a lifeform like KITT

KITT2kk · Post by **KITT2kk** » Fri Mar 28, 2003 4:40 am

Even though it is the 21st century, I still have thoughts that if anyone had money and technical support...Will it be possible to create an incredible lifeform such as K.I.T.T.?

knightimmortal · Post by **knightimmortal** » Fri Mar 28, 2003 12:47 pm

Not for the next 10-20 years at least.

(check board Search function for "AI")

KI

jup · Post by **jup** » Sun Mar 30, 2003 12:03 am

A better question is, 'Who is trying to make AI a reality, these days?'

I highly doubt that true computer intelligence can be met with today's silicon-based technology. (I've read that silicon technology has reached it's limits.) We need something else. Something with a far more advanced path structure that can process many times faster and RAM is never an issue with.

Does make me wonder if we will find it...or it will find us. I still see an off-branch of cloning as being the solution. If ever such a dangerous one.

CB2001 · Post by **CB2001** » Sun Mar 30, 2003 8:29 am

Actually, AI is already out in the world today. It's not just as advanced as KITT is. Where can you find these AIs? Simple, in first person shooters.

Post by **Michael Pajaro** » Sun Mar 30, 2003 2:55 pm

By that definition, AI also existed 30 years ago in early versions of Pong that you play against the computer. But a computer like KITT, that can carry out a true human-like conversation and is capable of real-world problem-solving is probably at least 50 years away if not an entire century.

Mike

knightimmortal · Post by **knightimmortal** » Sun Mar 30, 2003 6:37 pm

CB2001 wrote:Actually, AI is already out in the world today. It's not just as advanced as KITT is. Where can you find these AIs? Simple, in first person shooters.

That isn't even in the realm of AI, not even by definition. Sorry.

Mike- 10-20 years. (Providing we don't end up blowing ourselves back to the dark ages with the rest of our scientific endeavors.)

But yeah, I agree, the above definition pretty much puts Pong in that same criteria, which it most definitely does not.

KI

Post by **Michael Pajaro** » Sun Mar 30, 2003 7:05 pm

We'll just have to disagree. I still think my prediction of 50 years before seeing a KITT-like A.I. is wildly optimistic. I don't think we will see it during our lifetimes. We'll know in 20 years.

Mike

Army_F_Body · Post by **Army_F_Body** » Mon Mar 31, 2003 12:23 pm

I remember when I was in high school Sega released a game based on Jurrasic Park claiming the dinosaurs had true AI and would try to outsmart the player and learn your attack paterns. That game was good but the AI was a joke. I've never played an FPS where the enemies acted remotely intelligent. The games only have so many "patterns" programed into it that the programers try to say is AI. Most games only have two outcomes. One where the computer royally stomps you, they other were it just basically gives up and you win. One extreme or the other, 1 or 0. There seems to be no gray in between. Although I'll give it to you some of the next gen consule titles are pretty convincing, but it's probably just due to the fact that the storage media has more room to put the pre-programmed "patterns" into than the old days of cartridge gaming.

jup · Post by **jup** » Mon Mar 31, 2003 8:31 pm

I STILL like the errored way my ancient Commodore game, Lords of Conquest's AI worked. OK. The AI player would think and think and think. Suddenly, it would make a decision to obtain a property of mine. It would send in the troops and then the navy and armed forces. Problem was, it was breaking the rules and this second (honestly. A second) AI would pop in and tell the first AI "No. You can't do that." So, the first AI replies, "But, I wanna! O...Whatever." and it would return to thinking. Some of the time, it really did change it's mind. But, not always. This sort of AI player vs. AI unpire always cracks me up. (Well, I just mocked the conversation bit to help get the AI's split mind actions across.)

As for gaming AI, I would like to create one for my concept game, Street Mortality, that tries out maneuvers against the player as well as learns how to block successfully and use counter-attacks that the player isn't managing to defend against. I also want to blend it with a sim life engine in the back ground, so the game can self-generate the fighters and other people's life basics. A lot of randomness with a little bit of fuzzy logic in control can really generate some interesting, almost too intelligent results.

Rockatteer · Post by **Rockatteer** » Thu Apr 03, 2003 5:41 pm

I would have thought it was more the programming side of things which would let us down.

The technology we have right now is quite capable of supporting an AI system..assumeing your talking about Thought only, as in a computer which can think for itself. (if you wanted it to be mobile then you might have some problems

)

We have the technology, but we don't have the knowledge.

I don't recall ever hearing Kitt refered to as AI in any KR show. They always talked about software and hardwar...Kitts memory, cpu, circuit boards etc
The only thing that comes close is Bonnies comment to Micheal in "Junk yard Dog" where she says "Kitt was more the the sum of his parts"

Anyway thats my ten cents worth...I do beleive our current technology can supprot a thinking computer...but I don't think we know how to right the software to do it yet.

I guess the real question could be ...why would anyone want to?

Post by **Michael Pajaro** » Thu Apr 03, 2003 6:02 pm

I sort of feel the same way you do Rockateer... we don't need bigger or faster computers. With distributed systems technology, the processing power of computers is, for all practical purposes, limited only by the number of machines connected together. In theory, you could have a billion computers all churning away on the same problem and they would STILL have no greater "intelligence" than your GameBoy.

For a self-driving car, we do need computers fast enough to analyze their surroundings in real time at 60mph. And we can pretty much do that already. (but I think we're still 20 years away from having automated cars we can trust.) Like you say, until we understand what it means to think and to have intelligence, we can't even begin to translate that into ones and zeros for a computer.

Mike

knightimmortal · Post by **knightimmortal** » Thu Apr 03, 2003 8:34 pm

Now, let's bring in the knowledge, here shall we?

We have neither the technology or the software yet, or more specifically the hybrid that is needed.

For all you technical laypeople who have this media created view of AI, you sorta need to learn reality. We don't have computers that can support constant, and non-interrupted processing, but are working on it. We do not have circuit pathways that can handle that high amount of electrical input, to sustain the computational needs for it.

As for the software, a lot of it is already in existance, in parts and pieces as a whole for AI, and understanding what intelligence is, is not part of the problem. There are established tests, thought processes, and psychological analyses that give a base criteria, which is all that is needed for creation. After all, we only create what we know, but to get them in proper sync with the equipment to sustain it, that is the true boundary that exists, and is hard to cross.

So, for somebody who has worked in the field, compared to your average media influenced layman, no, we do not have the technology. We do not have the software, and we are right around 10-20 years from the first realization of what can fit all the criteria for TRUE AI.

KI

Rockatteer · Post by **Rockatteer** » Thu Apr 03, 2003 10:28 pm

OK just to pull you off your high horse.

Your not the only person in this forum with IT knowledge.

I'm not your average media influenced layman. I've worked in the guts of the IT industry for the past 12 years.. working with mainframe and super computers running huge database systems as well as minicomputers and PC systems using LAN and WAN running a host of cutting edge inhouse custome designed systems.

So perhapes a little less of the "high and mighty" attitude and a little more peer to peer conversation.

----------------------- now back to the topic --------------------------

Your saying that our current supercomputers don't have enough power to run a true AI application?

Or are you saying that a true AI system is more than a "thinking" computer and has to be some sort of android type self intity or somethign along those lines?

knightimmortal · Post by **knightimmortal** » Thu Apr 03, 2003 10:42 pm

Ok, I don't need you telling me to get off my high horse. I don't work in IT. I work in Computer Engineering with a basis in AI. I have worked in several different labs on several different development projects, so kindly take your lecture somewhere else. I wouldn't have peer to peer communication with somebody who so freely and publicly calls down somebody else as such to make yourself feel better. I am not some child, and I don't need you treating me as such. I wasn't referring to you as the layman, I wasn't referring to Mike as a layman, I didn't call you by name, so how about you get off your defensive end?

And as a matter of fact, no, the supercomputers don't have the power for sustained computational abilities, because they have to have lapse times, and true AI, can not have a lapse time. A true AI, (but what do I know, right?) is not just a thinking computer, but one that can emulate intelligence, and react to such. Not an android, no, but a computer that can do more than think, and can grow largely off of its own programming, not off of what people feed it.

KI

KITTFAN2003 · Post by **KITTFAN2003** » Fri Apr 04, 2003 12:33 am

Okay, I finally decided to take a shot at this one. Todays technology would be too slow and too dependent on constant programming to support a car like KITT. There is no technician anywhere that has near the training and know how that Bonnie supposedly did on KR.

Also, costs to build a KITT like car would be out of reach for even most companies, let alone people even if it were possible.

Finally, KITT or a car like him wouldn't be practical for the everyday driver. I mean, who needs turbo boost if they're just going shopping at Wal-mart

JMHO,

Post by **Michael Pajaro** » Fri Apr 04, 2003 1:24 pm

The greatest scientific minds alive can't agree on when or even IF machines will be able to think. So it would be arrogant of us to think that we could come up with a definitive A.I. timeline. The best we can do is post our opinions.

I'll argue that computers do not need to be fast to be intelligent. As a matter of convenience, we would like a smart computer to respond immediately if we ask it a question, but if it takes an hour to respond, or a day, or a week, so what? That might not be practical for most applications, but it doesn't mean that the machine isn't intelligent. We have to get away from the human-centric view of intelligence. If rational thought can be broken down into mathematical algorithms (which is a big philosophical debate), then today's computers could process them. Maybe not as quickly as we'd like, but it would still work. The problem is that we don't know how to translate human thought into binary.

Saying that we will have true A.I. in 20 years and saying that we could build KITT in 20 years are two wildly different claims. As far as AI goes, KITT is amazingly sophisticated. You can define Pong as being intelligent, or Google as intelligent, or a yet-to-be-built computer as intelligent. That's a question of definitions. But if you are specifically looking for a KITT-like computer, I'd wait 50 years, possibly 100.

Some people would say I'm pessimistic about AI, but I think I'm being realistic. History teaches us that we do a terrible job predicting the future of technology:

In 1930, people thought in 20 years skyways would replace roadways because everyone would be flying planes.
In 1950, people thought in 20 years atomic energy would solve all our problems and give us cheap, practically unlimited energy.
In 1970, people thought in 20 years we'd have a man on Mars and a hotel on the moon.
In 2003, when people say in 20 years we'll have cars driving themselves and sentient computers, my response is that history would say otherwise.

Mike

knightimmortal · Post by **knightimmortal** » Fri Apr 04, 2003 2:33 pm

I am not saying that the concept is speed, Mike, not at all. I am saying that the concept is the sustained long run to be able to continuously process code without a down period. We, as humans do not have a down period, even when we are sleeping, our brains are running, and that very concept, of keeping a computer constantly running at optimum efficiency is not possible. One lost link in a thought process would be like system failiure.

There are only a few HUGE problems with your vision of history saying otherwise:

1) Everybody envisioned wide scale replacement. That is not being done with AI.

Ex: Skyways replacing roadways. That was going to the far extreme. Previous to Sept. 2001, Skyways were as popular as roadways, as many people were flying world wide, as were on the roadways, on the average, and in fact, the antiquated rail and boat methods were scaled back drastically from what they were back in the 1930's, so in effect, that slowly became a reality, even within 20 years, the elevation of flight had become evident.

Ex: Nuclear energy has in fact given us cheap, practically unlimited energy, and could solve all our problems provided that people get past their fears of nuclear energy, but in fact, in a matter of 20 years, nuclear power plants became a fact of reality, and many cities and countries rely upon it. The mass change did not become a reality, just like all drastic changes haven't.

Ex: A man on Mars and a hotel on the moon. That was not necessarily a wide scientific belief. That was more science fiction fuelled intelligence. On the other hand, we have an international space station, which was in fact a scientific speculation back in the 70's, when the plans were laid down for such. So, that in effect was us right on schedule.

I didn't say at any point that we would have cars that drive themselves, have KITT, or such. I did say sentient computers, but I didn't say it would happen en masse. because I have seen what is going on in the R&D world of this particular subject.

KITT was a marvel, no doubt about it. One of the things that will be a hinderance on a sentient computer is the concept of enslaving a sentient being in a metal case with its only freedom being the internet. But in short, we will be capable of AI technology in 20 years, if we don't all end up blowing it straight up. And, if we ditch the grand added concepts that we keep adding on to it. No, not every computer will be sentient. No, not every computer will be in the form of a shiny black car that can drive itself. (Though once again, history does keep up with it's time line, there are already systems in place that allow a car to drive itself. It is the same concept as GPS auto pilot for jets.)

I'd say everybody thinks you are pessimistic about AI, Mike, almost to the point of being paranoid. Rather than supporting scientific reality, and scientific effort, you happen to always bring up the dark side, and don't inspire. Science is 99% Inspiration and 1% Persperation. Is it some personal offense to you for people to shoot for the gold and hold some hope in something, and work towards trying to achieve it? Or is it easier to bring your version of reality to the table to hold any hopes down?

History teaches only those who are willing to learn it correctly.

KI

Post by **Michael Pajaro** » Fri Apr 04, 2003 3:44 pm

I don't understand what computer downtime has to do with intelligence. How long would a computer have to be running before it would be considered reliable enough for AI? I routinely specify and maintain broadcast RAID arrays which go for years without any downtime. They churn data 24 hours a day, 7 days a week. In an emergency we can swap out a drive and have zero data loss. And there are for more reliable systems out there than what I'm using in the broadcast industry.

Why am I being paranoid just because I think sentient computers are still 50-100 years away? If you gave unlimited resources to Newton, DaVinci or Franklin, they would not have been able to go to the moon. I don't care how creative or inspired or intelligent they were, the world simply did not have the collective knowledge to do it. With unlimited resources, Einstein MIGHT have been able to do it. The same applies to AI. Even with unlimited resources, I don't think we could build KITT in 20 years. And there is no reason to think we could get unlimited resources in the first place.

Yes, cars and planes can have auto-pilots which work fine under CONTROLLED circumstances. But you have an automated car run into a detour sign and all of a sudden the car becomes useless. Until machines can adapt to the unexpected, they're not going to be truly intelligent.

I'm all for people having dreams, but they should also balance that with reality. No matter how determined or passionate I am about it, I'm not going to be the next American Idol. Sad but true.

If you want to give evidence supporting the idea of building a single sentient computer in 20 years, lay it on us.

And if you want to paraphrase Edison, he said "Genius is 99% perspiration, 1% inspiration". So even he seemed pretty pragmatic about progress.

Mike

knightimmortal · Post by **knightimmortal** » Fri Apr 04, 2003 4:10 pm

Exactly what evidence would you like, Mike? What evidence would soothe the savage pessimist, who has taken all articles and all examples of evolving AI technology and crapped all over them in the past. (Ref: the Popular Science article, which laid out that another key had been found in the path to AI.)

On an AI, you can't switch out drives. You can switch out memory modules, on a limited basis, and you still have to have redundant systems to pick up any missing time. How do I know? I have worked on projects for Intel, New Mexico Tech (the nifty little simulation program that I keep utilizing here that tracks with jup's Sim, is a bit of an example of it.) Stanford, and all of them are international projects that coincide with efforts overseas. I have played with those larger computers, and have seen what happens when a single data bit wipes out months of work.

An AI computer has to be able to constantly process all the time, even the supercomputers at Los Alamos can't do it, because they have data skips, it has to be processed constantly at a sustained consistent speed, and under those conditions even supercomputers start to have heatsink problems. So how about you stop treating me like I am some ignorant idiot who knows nothing? Just because I can't hand over the direct code for you to chew on, does not mean that things out there exist that you do not know about.

I happen to deal in reality as well. I also happen to deal with the concept that maybe I don't know everything, and have to accept it. I also have to deal with the fact that I have people like you who will never accept what I and others know, so it really isn't worth violating my security clearance for it.

And once again, Michael, READ: I didn't say we would have KITT in 20 years. NO KITT. Not KITT. KITT is a no. Negative. Zero. I just said that we will have AI in 20 years. I have said several times that it WOULDN'T be KITT.

We already have the basis for KITT working in a lot of the military systems, all of which have not been detailed for public consumption for security reasons. But it isn't full AI yet. It is the roadwork for such, which is more than you can say for most scientific discoveries which have been tripped on in the past. The same discoveries that in effect, brought fruition to sci-fi wants of the past.

And one other thing, the latest super computer in Los Alamos calculates the unexpected on its own. Is it intelligent? To a point. AI? Nope. Doesn't fit all the criteria. It knows when to switch its own systems off when there is a breach, or when there is a coolant problem. Another example: the Space Shuttle. It was trying to compensate for a problem, when nobody else was aware of such, it just couldn't compensate fast enough. And that is considering that the technology in it, even updated, wasn't as advanced as what some of the University and National Labs have going. So the signs are already here in the present day, but we still have a long way to go, in both hardware and software technology. But not 50-100 years. I'd give you 25 max, with the present development pace of technology. I could end up wrong, but not because of the evolution of technology going wrong, but because of the human element such as wars, terrorism, international relationship rifts, and pessimists going gung ho with their word en masse. Moralists will also be having their own fits.

Did you read any of this, or are you just going to argue your point about everything I say being fantasy again?

And it was Oppenheimer who rephrased it to apply to the proper scientific concept of 99% Inspiration, 1% Persperation. I hate to say it, but the more modern scientists seemed to have a little bit more of a grasp of technology and its applications rather than moralistic visions. Which is why today's scientists do more thinking and general theorizing than tripping on the inevitable.

KI

Post by **Michael Pajaro** » Fri Apr 04, 2003 8:34 pm

I accept your opinion that we could create a sentient computer with KITT-like intelligence in 20 years. (Not KITT specifically, just a computer as sophisticated as KITT). I just wish that you would accept my opinion that we won't.

Just because I'm posting a differing viewpoint doesn't mean I'm crapping over other opinions.

Just because I disagree with you doesn't mean I'm treating you like an ignorant idiot. (your words)

Just because I'm debating the issue doesn't mean I'm arguing against you personally.

Just because you think I'm paranoid, doesn't mean that I'm wrong.

Mike

knightimmortal · Post by **knightimmortal** » Fri Apr 04, 2003 8:39 pm

I would happily accept your opinion that we won't, if you didn't try once again to pull it off as fact.

The way you do, does.

No, but your commentary does about not realizing reality.

It sure seems like it is getting personal.

Because you are coming off as paranoid, makes one wonder where the ulterior motive is.

KI

Post by **Michael Pajaro** » Fri Apr 04, 2003 9:10 pm

I'd like to keep this from being a 2-person thread... what does the rest of the board think? What will it take to create a computer like KITT? Are we 5 years away? 25? Never?

knightimmortal · Post by **knightimmortal** » Fri Apr 04, 2003 9:14 pm

Mike, kindly do not try to be a moderator, and if you would like to keep it from becoming a two person thread, then don't continue to reply.

If others feel like jumping in, on the topic, they are free to.

KI

jup · Post by **jup** » Fri Apr 04, 2003 9:31 pm

Michael Pajaro wrote:In 1930, people thought in 20 years skyways would replace roadways because everyone would be flying planes.

The idea of flying a plane and the flying car was around back then and technically possible. But, the idea didn't really catch on with the public. Perhaps, the right advertising minds wern't employed on pushing the idea, or something.

In 1950, people thought in 20 years atomic energy would solve all our problems and give us cheap, practically unlimited energy.

Free, unlimited energy was a potential reality before the 20'th century. Again, the general public didn't sway to the notion enough to make it a reality.

In 1970, people thought in 20 years we'd have a man on Mars and a hotel on the moon.

This goes way back to an off-topic thread about Fox's Man on the Moon special. Which basically leads me to this way of thinking that until the technology of totally blocking out the deadly radiation levels in deep space are resolved, the Human race is STILL limited to looking, listening and sending out radiation-proof technology for it's space travel.

In 2003, when people say in 20 years we'll have cars driving themselves and sentient computers, my response is that history would say otherwise.

Mike

Again, we already have cars that can drive themselves. (Not that I would trust one to work 100% of the time.) Drive in any sense of the word "intelligently"? No. But, they can drive themselves...until something gets in the way and they wind up in a wreck.

As for what I expect the first AI's to be...

I believe that it will be a new form of OS. One that can understand the basics of English commands and speaks native computer language. One that NASA could use on their multi-billion dollar deep space probes so that it can properly explore and conduct on-the-fly alterations without requiring a Human's input or some sort of random factor to do a maneuver. One that could create a user interface for any foreseeable purpose and one that could put the computer programmer carreer on the endangered list. (Because, the only programming process left would be the 'wish commanding' process, which almost seems to be like a boss talking to a secretary while typing a letter.) But, even that AI wouldn't be anything near KITT's ideal levels. But, I wouldn't doubt the possibilities of seeing something like that emerge within the next 20 years.

Come to think of it, something like Star Trek: TNG's holodeck vocal interface comes close to what I envision. (Only, not so advanced.) Say, when a character walks in to that blank grid room and said character requests a piece of furniture and gives basic specs. Then, a piece of furniture appears. Next, the character redefines some of the furniture's looks and the computer resamples the furniture with the new info. Something like that is what I envision this new OS as being like.

Darknight · Post by **Darknight** » Fri Apr 04, 2003 11:13 pm

"The idea of flying a plane and the flying car was around back then and technically possible. "

Technically possible? Do tell. Flying cars/platforms have only been under serious development since the late 50s. Even now, the flying car concept is still very iffy. It relies solely on computers for stability, since it has no conventional controls surfaces...heck, it doesn't even have wings. Here's the latest versions of this idea.

http://www.skycar.com

"Free, unlimited energy was a potential reality before the 20'th century. Again, the general public didn't sway to the notion enough to make it a reality. "

I really hope you don't buy the "free energy" myths, even though there are some energy sources that are nearly free. There is no such thing as true free energy, though. No single energy solution exists. It will take a wide range of energy sources specially designed and adapted to fit certain locations.

AI is a very delicate subject. Most individuals do not even know the right questions to ask, much less the answers to them. If AI is ever made a reality, and proven to the masses, it will require drastically different approach than simply making faster computers, and writing more complex programs. Digital technology is fundamentally flawed (and incapable of supporting AI by itself), is all I'm gonna say;)

DK

Do you think our technology can support a lifeform like KITT

Do you think our technology can support a lifeform like KITT

knight rider: original

knight rider: 2008-2009

team knight rider

other knight rider

news

user contributed

videos

photos

follow us