Not only do they shoot, they appear to try to arrest people, if I'm interpreting the woman raising her hands in response to what sounds like a recording coming from the robot correctly. They don't appear to be mobile; more for "sentry" applications.
Although I keep up with Robotic programs at MIT and elsewhere I'm hardly suprised to see the Market well ahead in many respects of governmental, military R&D, and academic efforts.
Coincidentally enough I recently wrote an article about matters related to these developments. And I'm also working on an operating system called HUMAN.
Of course this now means that when it comes to total war, the age for eligible draftees can now span 14-85, with those not of prime ground-pounder age sitting back and handling the remote robots. Basic vision, hearing, and a low level of mobility will suffice to control the robots.
I wasn't impressed, though, with the speed of swiveling the turret nor with the elevation for AA use. I think it could use a little tweaking there. The demo seemed to be a 7.62/.30 cal. machine gun. Do they make them in larger sizes? That'd be awesome in both .50 cal and 20/25mm!
As I get older, I disagree more and more with the sentiment that a human needs to be in the loop with something like this.
My engineering experience (and years of data) has shown me that anything that can be done by machines, probably should be done by machines. Machines are so much more predictable. And in fact, they are the only hope for ever achieving a failure rate of zero.
In manufacturing, we can develop automated processes today with defect rates that are practically un-measurable. But add one human step to any process, and imperfection is guaranteed, eventually.
It’s more difficult for us to accept this where human life is concerned. But we already have turned control over to robots in many areas where it once seemed unthinkable: flying a plane, regulating trains, controlling traffic, docking a ship, etc. They do a better job, on average, then people.
And in reality, the same process is already at work with most smart weapons we use today.
So I have less qualms about an armed robot (properly designed &tested), than I do about an armed human security guard.
I think a properly designed and tested autonomous robot is probably safer than one with a human operating it. Even setting aside human error, a human operator can be incapacitated, bribed, blackmailed, threatened, etc. The control mechanism the legit operator uses to operate the robot can probably be compromised, too. That would be bad.
I'm with Bryan for the most part, although giving a robot a gun still manages to disturb me.
Yet I think it's an inevitability that we're going to see robots in more and more of these roles as artificial intelligence continues to improve. I'm most looking forward to the self-driving cars myself. I doubt we'll have to wait even ten more years for that.
Well, if you deploy these robots along the demilitarized zone, what's to screw up? Their job is to shoot everything that even vaguely resembles something that can move until there's no longer any chance that it can move.
Plus, this has the advantage of being able to outlast the humans, thus leaving a legacy of automatic doomsday machines which whatever alien race eventually lands here can then have to escape from.
I think that such robots are really cool. On the other hand, they better not have network-upgradeable firmware. Then all it takes is one guy to sleep with the wrong "person" and suddenly all of our robot defenders turn on us. (Or if you'd prefer a B5 reference, one untrustworthy politician and you get operation scorched earth.)
I dearly hope that you're right about auto piloted cars (I think that we could nearly have them now if all you want is highway autopilot), but I think that you're way off. Do you really think that auto piloted cars will ever make it past the car manufacturers legal department in the next decade?
Hmm... my atavistic, gut-level reaction is that, even if we normally trust robots to handle the weapons autonomously, a well-trained human needs to be Oversight... not to constantly give the order to fire, but in case a fail-safe cut-out is needed.
Now, sentry situations like the DMZ? No problems with it.
Don't think that people a ready for this kind of things,it's like about "Lie Detector" ,although it can tell if somebody lies with a great precision yet it isn't used in The Justice System.But I'm all for these doing all the routine operations,or most dangerous things like detecting IED's .Bad thing however is that once becomeing very accesible it will be used by insurgents as well , for exemple mobile phones used as detonators,or unmaned planes as a bomb(Attack on Israeli warship),Katiusha rockets fired remotely,MassMedia.It's for technology that in this world terrorism is so dangerous,thats the biggest problem,I don't want to see a world where a bunch of crooks can terrify everybody,or elect presidents/prim_ministers(like Spain),so don't make these things to cheap and accesible.
What they'll obviously do is put the tech first in their high-end, expensive models, and after working it out with their insurers they'll pad in enough to deal with liability in the initial models. Then as the technology continues to improve and problems are worked out it'll become available in more common and cheaper cars.
Yeah I'm saying this will be pretty common technology in a few years.
Jaymaster: I don't doubt for a moment that machines make fewer errors than humans, if programmed correctly. That's not the issue at stake here, though.
What does matter is what happens when someone is wrongly killed. Suing the manufacturer (and howmany other deep pockets one can find) doesn't really do it. People want a human to take responsibility--and pay the consequences--of a screw-up. They don't need to go off all fretting about "Skynet" to be concerned that a machine, despite excellent programming, will come across a situation outside its program's bounds and not do what a human would do, i.e. "the right thing".
Leaving such a machine on autopilot, even in a DMZ, is not safe from catastrophic failure. Will the machine's sensors distinguish a refugee or deserter trying to get to the other side? Shooting refugees is really bad PR; shooting deserters can be extremely expensive intell-wise. Both tend to discourage others from doing the same, even using different ways.
No one was actually ready to hand the keys over to a "Doomsday Machine" that would launch-on-launch. We and the Soviets all wanted a human--fragile as s/he may be--to be a major component of the loop. It's all about responsibility and accepting it.
So until AI shows a competent awareness of self and other, I'm not going to be voting to downsize the military....
No one was actually ready to hand the keys over to a "Doomsday Machine" that would launch-on-launch. We and the Soviets all wanted a human--fragile as s/he may be--to be a major component of the loop. It's all about responsibility and accepting it.
And as most of us know, that choice probably saved us all from WW3. Remember this guy?
http://en.wikipedia.org/wiki/Stanislav_Petrov
If that had been a robot or a computer chances are we'd all be dead.
BK: I'd forgotten about the self-parking cars. Yep, we already have those.
Jesse: Given that I just noted that General Motors has announced that they'll have a self-driving option on the market by 2008, what do you base your assertion on? That they lied? That no one will buy it?
Some people want control. Me? I'd love to read a book or, better yet, surf the internet while my car drove itself.
Come to think of it the problem of drunk drivers would also be greatly reduced here.
Given that we now know the technology is working and will be on the market soon, I actually find the much more interesting question to be, "how long before driving your own car is illegal except in emergencies?" My guess: before the end of the 2020s.
I think that the most necessary ingredients for production are
Land
Raw materials
Labor
Energy
Knowledge
I will venture to guess that when we figure out how to harness the practically limitless energy readily found in our environment,land will not be scarce,as transportation will becomes inexpensive, and raw materials will not be scarce either,as we will be able to look for them at depths that were hitherto uneconomical to excavate.Labor is increasingly plentiful,and robotics will all but make it moot.
We will live in an era of unimaginable plenty,man
will have to find better uses for his time.
I've been mulling over this for quite some time.Any thoughts?
Such an explosive laden vehicle, disguised as a normal auto could also be used to maneuver near or crash into and destroy security robots of the kind described by the original article.
Such a car would be a tactical level fire-ship.
Then again it could also be deployed by terrorists as a conductor for thermobaric explosives, as per this article: http://themissal.blogspot.com/2006/11/new-weapon-plot.html
So the uses could potentially be quite varied, depending upon what type of operation were contemplated and who was undertaking the operation.
You know I have recently been considering the idea of developing a Non-Lethal Weapon's Vest, similar in design and disguise capabilities of the Homicide/Suicide Vest of terrorists. But in this case the vest would be worn by a volunteer, solider, detective, Hostage Rescue agent, soldier, whose job it is to infiltrate and penetrate a dangerous situation or area and then discharge his vest. The vest would contain an area suppression weapon which would incapacitate or render unconscious everyone within the tactical perimeter of the attack, or within the effective range of the Non-Lethal or Less Than Lethal system employed.
Of course I would prefer to see new and far more effective NL and LTL systems developed and designed for use in such Vests than we currently have in use today.
I dearly hope that you're right. You have no idea how desperately I want a car that can drive itself on the highways.
But I'm way more pessimistic than you are about the legal situation. It's not just the manufacturer. How many insurers will insure people with a self-driving car? Think about the personal lawsuits if you are using the self driving option, read a book, and your car hits someone else. What if there was a malfunction, and your car kills several people. Will you be criminally libel?
There are a lot of issues to self driving cars, and I just don't have your faith that they're going to be worked out in a mere 10 years.
And then how long until it's available to people who only buy sub-$20,000 cars (like me)?
Naftali: A very good book to read on this is Ray Kurzweil's The Singularity Is Near. Which deals with a lot of these issues. Yes, we are rapidly coming up on a world in which there will be no more manual labor jobs. Amongst other things. The social changes from that are going to be even more massive than the industrial revolution.
Chris: Someone like General Motors is not going to put a car like this on the road unless they feel extremely confident and have already worked out liability concerns.
You ask what happens if your personal self-driving car gets into an accident? Initially, the law will hold that you Chris Lansdown the car's owner and operator is responsible. You'll have the immediate liability regardless. But then maybe you'll turn around and sue GM, but GM will in the meanwhile have locked most of this up in the purchase agreement/warranty statement, and further will have already socked away money to deal with possible lawsuits from unexpected failures.
Long-term there's simply no question that a self-driving car should be orders of magnitude safer than human drivers.
As for it being cheap: surely you must have noticed by now that most technology gets cheaper and most plentiful over time? It usually hits this pattern:
1) Is very expensive, is only used by a few individuals, and only works moderately well.
2) Is somewhat less expensive, and somewhat more reliable, and is used by quite a few in dividuals.
3) Is fairly inexpensive and most people have access to it, and it works better than ever.
4) Is dirt cheap and extremely reliable and just about everyone has it.
Every major technological breakthrough I can think of in my lifetime has gone through all four of those stages.
The self-driving car will first be a very expensive option that is only offered on a few high-end very expensive automobiles. But the drive toward cheapness and ubiquity will be continual.
"But the drive toward cheapness and ubiquity will be continual."
Sure. But how long will it take?
Personal helicopters in our backyards still aren't here. Space travel has had decades to drop in cost to where it's as affordable as a vacation to the grand canyon. It still isn't.
Adaptive cruise control (you know, where it slows down if the guy in front of you isn't going fast enough) still isn't standard on all cars. Hell, air conditioning still isn't standard on all cars.
And you know, I'm not very comfortable with the idea that my automatic car screws up and I go to jail for manslaughter.
You're right that it's encouraging that GM is planning to roll this out. Then again, The Coca Cola Company planned a big roll out for New Coke. Sometimes companies find out that great ideas don't work, and roll them back.
Bottom line: I'd really like some better evidence than you being an incurable optimist. :)
Because the general trends making auto-pilot cars likely isn't good enough. I want an auto-pilot car more than any other technological and social break-through. I don't want this to go the route of uranium powered cars — a great idea that never materializes.
We already have hydrogen cars. The problem is that hydrogen is an inefficient fuel source and costs too much.
As for my evidence being based on "incurable optimism"--bollox, I've already demonostrated several hard facts on the matter, including the simple reality that the world's largest car manufacturer has made it clear they're going to introduce this and that other companies are clearly working on it in a very serious way.
As for personal helicopters or space travel: personal flight will never become a practical reality until we basically have computers doing the flying. Because we could all have such vehicles now, but they're extremely dangerous even in the hands of experienced pilots.
But this gets us back to, especially, the simple iron laws of technical development, in the area of computing and AI especially. AI and robotics are what you really want to be looking at here, and the tremendous strides being made in those fields, with no end in sight; Moore's Law is still active here and in several other fields.
Still, I can't convince someone who is convinced that they need to be pessimistic.
I was asking for additional evidence, not recycling the evidence already presented.
And you're sidestepping the issue of hydrogen cars entirely. I know that Honda makes the FCX. The point is that I can't buy one at my local honda dealership for less than $1000. Technological progress should make hydrogen an efficient fuel source which is virtually free.
Technological progress inexorably makes things cheaper and more available, but not always at a convenient pace. If you want to claim that self driving cars will actually work out in the marketplace some time before the heat death of the universe, I'm with you. It's the next several years thing.
As I said, it's impressive that GM is making a bold move. But companies make bold moves and pull out. Or go bankrupt. (If the company goes bankrupt, it's generally not called a bold move any more, it's now a stupid move.) It's a good sign that GM is trying this out in their luxury cars.
But you're claiming, or seem to be tantalizing me with the idea, that auto drivign is going to become cheaper and more ubiquitous than air conditioning in the next few years.
I remember back when Toyota was going to have their entire line be hybrids-only by 2008 (and due to the volume, there'd be no price premium for hydrids). It's not looking very good for that to happen now.
Dean, you have no idea how much I want auto-piloting in sub-$20,000 cars. But can you name any advanced feature which went from high end luxury cars to near-ubiquity in less than 10 years?
That's encouraging (though XM receivers aren't a very advanced technology, so far as I know).
Still, Dean is promising auto-piloting cars in the sub-$20,000 within 10 years. I desperately want to believe him, but this seems like a stretch to me.
(And this actually effects me; my current car is starting to get older. The question coming up is do I replace it in 3 years or hang on and replace it in 6 or 7 or 8 or... The last thing that I want to do is get the last model before auto piloting is standard, and have the longest possible wait until I can get an auto-piloting car.)
And just in time for Christmas too.
Coincidentally enough I recently wrote an article about matters related to these developments. And I'm also working on an operating system called HUMAN.
God Technology
Of course this now means that when it comes to total war, the age for eligible draftees can now span 14-85, with those not of prime ground-pounder age sitting back and handling the remote robots. Basic vision, hearing, and a low level of mobility will suffice to control the robots.
I wasn't impressed, though, with the speed of swiveling the turret nor with the elevation for AA use. I think it could use a little tweaking there. The demo seemed to be a 7.62/.30 cal. machine gun. Do they make them in larger sizes? That'd be awesome in both .50 cal and 20/25mm!
As I get older, I disagree more and more with the sentiment that a human needs to be in the loop with something like this.
My engineering experience (and years of data) has shown me that anything that can be done by machines, probably should be done by machines. Machines are so much more predictable. And in fact, they are the only hope for ever achieving a failure rate of zero.
In manufacturing, we can develop automated processes today with defect rates that are practically un-measurable. But add one human step to any process, and imperfection is guaranteed, eventually.
It’s more difficult for us to accept this where human life is concerned. But we already have turned control over to robots in many areas where it once seemed unthinkable: flying a plane, regulating trains, controlling traffic, docking a ship, etc. They do a better job, on average, then people.
And in reality, the same process is already at work with most smart weapons we use today.
So I have less qualms about an armed robot (properly designed &tested), than I do about an armed human security guard.
Yet I think it's an inevitability that we're going to see robots in more and more of these roles as artificial intelligence continues to improve. I'm most looking forward to the self-driving cars myself. I doubt we'll have to wait even ten more years for that.
Plus, this has the advantage of being able to outlast the humans, thus leaving a legacy of automatic doomsday machines which whatever alien race eventually lands here can then have to escape from.
I think that such robots are really cool. On the other hand, they better not have network-upgradeable firmware. Then all it takes is one guy to sleep with the wrong "person" and suddenly all of our robot defenders turn on us. (Or if you'd prefer a B5 reference, one untrustworthy politician and you get operation scorched earth.)
I dearly hope that you're right about auto piloted cars (I think that we could nearly have them now if all you want is highway autopilot), but I think that you're way off. Do you really think that auto piloted cars will ever make it past the car manufacturers legal department in the next decade?
Now, sentry situations like the DMZ? No problems with it.
Would make a good explosive-laden car.
GM Will Launch Self-Driving Car In 2008.
Volkswagen is also close.
What they'll obviously do is put the tech first in their high-end, expensive models, and after working it out with their insurers they'll pad in enough to deal with liability in the initial models. Then as the technology continues to improve and problems are worked out it'll become available in more common and cheaper cars.
Yeah I'm saying this will be pretty common technology in a few years.
What does matter is what happens when someone is wrongly killed. Suing the manufacturer (and howmany other deep pockets one can find) doesn't really do it. People want a human to take responsibility--and pay the consequences--of a screw-up. They don't need to go off all fretting about "Skynet" to be concerned that a machine, despite excellent programming, will come across a situation outside its program's bounds and not do what a human would do, i.e. "the right thing".
Leaving such a machine on autopilot, even in a DMZ, is not safe from catastrophic failure. Will the machine's sensors distinguish a refugee or deserter trying to get to the other side? Shooting refugees is really bad PR; shooting deserters can be extremely expensive intell-wise. Both tend to discourage others from doing the same, even using different ways.
No one was actually ready to hand the keys over to a "Doomsday Machine" that would launch-on-launch. We and the Soviets all wanted a human--fragile as s/he may be--to be a major component of the loop. It's all about responsibility and accepting it.
So until AI shows a competent awareness of self and other, I'm not going to be voting to downsize the military....
And as most of us know, that choice probably saved us all from WW3. Remember this guy?
http://en.wikipedia.org/wiki/Stanislav_Petrov
If that had been a robot or a computer chances are we'd all be dead.
Jesse: Given that I just noted that General Motors has announced that they'll have a self-driving option on the market by 2008, what do you base your assertion on? That they lied? That no one will buy it?
Some people want control. Me? I'd love to read a book or, better yet, surf the internet while my car drove itself.
Come to think of it the problem of drunk drivers would also be greatly reduced here.
Given that we now know the technology is working and will be on the market soon, I actually find the much more interesting question to be, "how long before driving your own car is illegal except in emergencies?" My guess: before the end of the 2020s.
Land
Raw materials
Labor
Energy
Knowledge
I will venture to guess that when we figure out how to harness the practically limitless energy readily found in our environment,land will not be scarce,as transportation will becomes inexpensive, and raw materials will not be scarce either,as we will be able to look for them at depths that were hitherto uneconomical to excavate.Labor is increasingly plentiful,and robotics will all but make it moot.
We will live in an era of unimaginable plenty,man
will have to find better uses for his time.
I've been mulling over this for quite some time.Any thoughts?
He jumped on the keyboard as I was reading through the thread, and somehow managed to hit the “post” button as I chased him off.
Is that a first for Dean’s World?
Now that I think about it, such an act could actually have some significance with this topic…
A Fire-Ship.
Such an explosive laden vehicle, disguised as a normal auto could also be used to maneuver near or crash into and destroy security robots of the kind described by the original article.
Such a car would be a tactical level fire-ship.
Then again it could also be deployed by terrorists as a conductor for thermobaric explosives, as per this article: http://themissal.blogspot.com/2006/11/new-weapon-plot.html
So the uses could potentially be quite varied, depending upon what type of operation were contemplated and who was undertaking the operation.
You know I have recently been considering the idea of developing a Non-Lethal Weapon's Vest, similar in design and disguise capabilities of the Homicide/Suicide Vest of terrorists. But in this case the vest would be worn by a volunteer, solider, detective, Hostage Rescue agent, soldier, whose job it is to infiltrate and penetrate a dangerous situation or area and then discharge his vest. The vest would contain an area suppression weapon which would incapacitate or render unconscious everyone within the tactical perimeter of the attack, or within the effective range of the Non-Lethal or Less Than Lethal system employed.
Of course I would prefer to see new and far more effective NL and LTL systems developed and designed for use in such Vests than we currently have in use today.
Jack
I dearly hope that you're right. You have no idea how desperately I want a car that can drive itself on the highways.
But I'm way more pessimistic than you are about the legal situation. It's not just the manufacturer. How many insurers will insure people with a self-driving car? Think about the personal lawsuits if you are using the self driving option, read a book, and your car hits someone else. What if there was a malfunction, and your car kills several people. Will you be criminally libel?
There are a lot of issues to self driving cars, and I just don't have your faith that they're going to be worked out in a mere 10 years.
And then how long until it's available to people who only buy sub-$20,000 cars (like me)?
Chris: Someone like General Motors is not going to put a car like this on the road unless they feel extremely confident and have already worked out liability concerns.
You ask what happens if your personal self-driving car gets into an accident? Initially, the law will hold that you Chris Lansdown the car's owner and operator is responsible. You'll have the immediate liability regardless. But then maybe you'll turn around and sue GM, but GM will in the meanwhile have locked most of this up in the purchase agreement/warranty statement, and further will have already socked away money to deal with possible lawsuits from unexpected failures.
Long-term there's simply no question that a self-driving car should be orders of magnitude safer than human drivers.
As for it being cheap: surely you must have noticed by now that most technology gets cheaper and most plentiful over time? It usually hits this pattern:
1) Is very expensive, is only used by a few individuals, and only works moderately well.
2) Is somewhat less expensive, and somewhat more reliable, and is used by quite a few in dividuals.
3) Is fairly inexpensive and most people have access to it, and it works better than ever.
4) Is dirt cheap and extremely reliable and just about everyone has it.
Every major technological breakthrough I can think of in my lifetime has gone through all four of those stages.
The self-driving car will first be a very expensive option that is only offered on a few high-end very expensive automobiles. But the drive toward cheapness and ubiquity will be continual.
Sure. But how long will it take?
Personal helicopters in our backyards still aren't here. Space travel has had decades to drop in cost to where it's as affordable as a vacation to the grand canyon. It still isn't.
Adaptive cruise control (you know, where it slows down if the guy in front of you isn't going fast enough) still isn't standard on all cars. Hell, air conditioning still isn't standard on all cars.
And you know, I'm not very comfortable with the idea that my automatic car screws up and I go to jail for manslaughter.
You're right that it's encouraging that GM is planning to roll this out. Then again, The Coca Cola Company planned a big roll out for New Coke. Sometimes companies find out that great ideas don't work, and roll them back.
Bottom line: I'd really like some better evidence than you being an incurable optimist. :)
Because the general trends making auto-pilot cars likely isn't good enough. I want an auto-pilot car more than any other technological and social break-through. I don't want this to go the route of uranium powered cars — a great idea that never materializes.
And when are we going to get hydrogen cars?
As for my evidence being based on "incurable optimism"--bollox, I've already demonostrated several hard facts on the matter, including the simple reality that the world's largest car manufacturer has made it clear they're going to introduce this and that other companies are clearly working on it in a very serious way.
As for personal helicopters or space travel: personal flight will never become a practical reality until we basically have computers doing the flying. Because we could all have such vehicles now, but they're extremely dangerous even in the hands of experienced pilots.
But this gets us back to, especially, the simple iron laws of technical development, in the area of computing and AI especially. AI and robotics are what you really want to be looking at here, and the tremendous strides being made in those fields, with no end in sight; Moore's Law is still active here and in several other fields.
Still, I can't convince someone who is convinced that they need to be pessimistic.
On the other hand, I'm not sure the guards on the other side of that border are a whole lot less robotic.
I was asking for additional evidence, not recycling the evidence already presented.
And you're sidestepping the issue of hydrogen cars entirely. I know that Honda makes the FCX. The point is that I can't buy one at my local honda dealership for less than $1000. Technological progress should make hydrogen an efficient fuel source which is virtually free.
Technological progress inexorably makes things cheaper and more available, but not always at a convenient pace. If you want to claim that self driving cars will actually work out in the marketplace some time before the heat death of the universe, I'm with you. It's the next several years thing.
As I said, it's impressive that GM is making a bold move. But companies make bold moves and pull out. Or go bankrupt. (If the company goes bankrupt, it's generally not called a bold move any more, it's now a stupid move.) It's a good sign that GM is trying this out in their luxury cars.
But you're claiming, or seem to be tantalizing me with the idea, that auto drivign is going to become cheaper and more ubiquitous than air conditioning in the next few years.
I remember back when Toyota was going to have their entire line be hybrids-only by 2008 (and due to the volume, there'd be no price premium for hydrids). It's not looking very good for that to happen now.
Dean, you have no idea how much I want auto-piloting in sub-$20,000 cars. But can you name any advanced feature which went from high end luxury cars to near-ubiquity in less than 10 years?
Satellite navigation is well on its way. You can get it in a Camry now; it'll be in everything with the next model refresh, I wager.
I believe Hyundai has announced it's going to be making XM satellite radio receivers standard in all of its cars.
GM's OnStar is trickling down to the lower-end cars, you can get it in a Malibu now.
We're probably not in the "under 10 years" timeframe for any of those, but we're damn close.
That's encouraging (though XM receivers aren't a very advanced technology, so far as I know).
Still, Dean is promising auto-piloting cars in the sub-$20,000 within 10 years. I desperately want to believe him, but this seems like a stretch to me.
(And this actually effects me; my current car is starting to get older. The question coming up is do I replace it in 3 years or hang on and replace it in 6 or 7 or 8 or... The last thing that I want to do is get the last model before auto piloting is standard, and have the longest possible wait until I can get an auto-piloting car.)