Slashdot Mirror
Japan's Proposed 30-Year Robot Program
Gallamine writes "A group of Japanese researchers have proposed a Government plan to spend 50 billion yen per year (that's over 400 million $US) for 30 years on developing a robot with capabilities of a 5-year-old. Japan's current economy may prevent the plan from happening, but the interesting point is the parallels to the U.S. Apollo space program, America's attempt to put a man on the moon. While expensive, the benefits to the American population from that program are probably unmeasurable. Perhaps the U.S. Government should consider funding such a program over here?"
Perhaps instead, the US government should stop cutting funds allocated to education and "liberating" oil-producing countires.
Lets worry about the robots after we figure out how to pay back our debt.
10% Cute (or ugly to eveyone other than owner)
40% Crying
5% Crayon ability
15% Get daddy a beer
7% Underfoot
3% Questions beginning with 'Why'
20% Screaming, running, and breaking.
Please contact me for licensing.
let's let Japan do all the spending on the project, then we'll buy one of their fancy schancy new robots, and reproduce it ourselves.
I'm much rather have a 5 year-old with the capabilities of a robot.
Best Windows Freeware
Does the use of '5 year old boy' as opposed to '5 year old girl' make anyone else a little uneasy? Actually thinking about it neither is particularly suited to the hotbed of hormones that is slashdot. Why not say 'equivalent to an average windows user'?
If you double the price and can get me a robot with all the capabilities of a 19 year old cheerleader, I'll call my congressman tonight.
I knew Japanese Anime are more of a documentary then entertainment! I can't wait to see huge robots fighting each other, being able to transform into jets and guardian modes! Plus with all the destruction that the robots will make, the Japanese construction companies will be busy for quite some time!
We don't even have a Big Guy to go with him!
"Lord, grant that I may always be right, for Thou knowest that I am hard to turn" -- A Scots-Irish prayer
I have to wonder how they define a 5 year old and to what context? Ability to learn, reason, think interact. Also what about physical appeareance are they looking for same size or something the size of a fully grown human?
Could it get to the point where you have a "child" in a super human body? Hopefully they will have Asimovs Rules in there at least
Rus
Cheap UK and US VPS
This isn't the first time Japan is doing one of these long term plans. I watched a program a few years back explaining that japan had several plans like this ("tommorow's world" for people in the UK). Firstly they did a huge investment into transistors then silicon manufacturing and at the time of the program (1995 ish) they were part way through a huge investment into flat screen displays (not even TFT at that stage I dont think).
At the time I was thinking it was a huge mistake. Flat panes were slow, small and hugely expensive and no one would spend extra to have one to replace a better CRT. Im sure people were thinking the same sort of things on the other projects but they sure did pay off.
I'm not sure how Japan figures out what to pick but it seems to work. Maybe they are making very good choices or maybe if you stick enough money into something it will eventually pay off. And as sceptical I am of humanoid robots I can't say this is a silly idea any more.
Mouse powered Chips, Open source Processors and Lego
Heh, maybe it could work for Microsoft.
I'm a bit uncomfortable with the concept of building life, even artificial life such as this. Who is to say that we would hold the same degree of ethical concern for something we can build in a factory as we do for life, or that these robot lives won't be held in less esteem than our own? Perhaps we should leave playing God in the hands of our creator?
Just in time to send it to Mars to work on the power plant.
It would probably take my girlfriend and I about four years to produce something with the capabilities of the average 5-year-old.
I'm pretty bright, and my girlfriend recently graduated from CMU with a degree in CS, and is now attending Johns Hopkins. It would (roughly speaking) take a 4-year-old child with an IQ of 125 to match a 5-year-old.
And for the quarter billion per year Japan is spending, I'd be able to afford some pretty neat educational toys, too!
Somebody get that guy an ambulance!
Way back when ('80s i think) the Japanese Government (MITI) decided that the Prologue language (Proudly Recursive - great for the Tower of Hanoi problem) was to be the basis for the artificial intelligence revolution that they wanted to lead. Don't remember? The problem with planning for the future is it keeps changing.
"Knowing everything doesn't help..."
The difference being that we can see the moon. Since we can see it, it was easy to suppose that there must be a way to get there if we only tried hard enough. But in the field of AI it's quite a different story. AI research is 60 years old now and the best we can do is emulate the intelligence of a cockroach. We're nowhere near building a robot that can simulate the intelligence of a 5 year old human. In fact, it may very well be impossible. (Personally, I believe that an intelligent digital computer will never be possible.) If that is the case then spending millions or even billions won't make much of a difference. Granted, even if the project fails a lot of new and useful technology will be created in the process. But overall I think I'd be more excited to hear that the US is going to commit to landing on Mars by 200x.
Support the First Amendment. Read at -1
Japan has pitched an idea to create in 30 years a a robot with the capacity of a 5 year old child. The idea struck a cord in me, and I decided to take a break and just think on that.
Let's define the nature of the ultimate goal in robots in the business - I'm sure there will be quibbles, but this is my definition:
Ultimate Goal: To create a robot with human level intelligence and physical manipulation without sentience or self awareness.
By this definition, we mean the capacity to learn, to be instructed in tasks and incorporate ideas into itself and understand commands without detail, but without sentience or self awareness, never having emotions or being able to make fully independant decisions about freedom, what to do for itself.
In a word, the ultimate metal slave.
Let's throw ethics out the window for a moment - we'll get to those in a minute. But let's say you could make such a machine. One that you could give orders to "go clean the house", and it would intelligently understand and fulfill your wish without the "evil genie" effect (where a badly ruled wish has unintended consequences - see "The Monkey's Paw" for an example, where you could wish for a million dollars, and you would get it - after your son was killed in an automobile accident and the money was payment from a life insurance policy).
Ignoring if such a goal is possible (and, seeing how far we've come in 100 years, is it so far to reach that in 500 years we would be capable of building such a machine?), let's see what would happen to society.
Employees, especially blue collar, farmers, manufacturing and the like, could be mass produced. A whole army of robots that would work without tire, without pay, and if you could make them mass produced to be cheap (say $20,000 - $40,000 a year), if they break, get a new one. They could work day and night, rotating in 8-12 hour shifts for maintenance and repairs. Farms could be worked all day long, and if there was a problem, robots could go out and fix the issue. Need to pick the cotton/coffee beans? Just hire the robots to go out and do it. Wars fought by machines - never tiring, truly "bloodless" wars where a million "soldiers" could be airdropped into the field loaded with advanced weapons to wipe out the enemy by beings that have no conscience. (Granted, hacking would truly become the greatest weapon in society at that point, but just go with me a moment on the idea.)
Food prices, car prices - hell, prices for everything could actually drop, since the human cost of making them would be negligable. Ah - but for one major problem:
What do the people do?
Millions - let's even say 25% of the work force alone, just to argue - out of work. They're not needed at McDonald's or Ford or even Dell - replaced by machines. So what do they do? Not everybody could work in a robot making factory. Does the world start to become a place where human labor is practically no longer required? Where only a few work because they want to to design new things or create art, while millions simply live a life of leisure? Where everyone is guarunteed a certain level of life and comfort, and those who want more can sell their services of entertainment or some unique idea they are able to create in this new utopia of fully attained basic life for all people?
Or a world where millions can not get work and search but become homeless? If people think that having jobs from their country exported to foreign places willing to do it for less, how will they feel when the factory is still on native soil, but the jobs are for those tireless, non-paid, non-complaining machines? When they can't provide for their children, and the line between "haves" and "have-nots" is larger than ever?
I actually see a lot of promise in the idea - I really do. The benefits to business, to humanity could be huge. But I have the feeling should such a creation actua
52 Weeks, 52 Religions with John Hummel
I must ponder this question: Before we create machines with human-level intelligence, shouldn't we first ask "why?".
As it is, we're running out of human jobs to do (McDonald's for example is toying with the idea of fully-automated vending machines), so what will happen when we can make machines that can work for almost nothing, and start replacing human jobs? And what will happen if and when these machines start thinking by themselves (in which case they will demand rights, just as we do) and if they decide that they don't need us?
I'm not saying we shouldn't do it, I just think we should be careful on _how_ we do it. I'm actually a believer that at some distant point in the future we humans will slowly evolve into machines, and _then_ at that point creating more machines will be a natural thing for us.
So I must ask, should we spend all those billions on machines instead of education? I don't want to sound like a miss universe contestant but right now world peace, world hunger, and world education should be our top priorities.
Again, don't bash me, I'm a true geek, I love machines, robots, AI, etc, it's just that I think we should spend some time thinking about the big issues facing humanity today.
On a related side note, space exploration is probably where I see the best use for robots.
Two words: Slave Labor. Robots don't threaten to walk out if working conditions are dangerous.
Don't trust any concentration of power.
Don't get me wrong, I still love research for research's sake, but I think that Japan could better spend $50 billion on more advanced urban architecture and transportation.
Why don't you go visit Japan? That way you will have a decent understanding of the way life is both in Tokyo and outside of Tokyo.
Dacels Jewelers can't be trusted.
I 've a couple. Will trade in for half the money.
New year Resolution: Don't change sig this year
I googled for some information and came up with this site.
Here are some relevant quotes:
Housing is a particularly severe problem. The quality built into a modern home is good, but cost and size are a different matter. With the rapid appreciation of land values in the 1980s in Japan, a four-bedroom, ranch-style, North American house on an acre of land and valued at maybe US$250,000 would have cost many millions in Tokyo. The most that people could hope for would be a 3DK (3 bedrooms, a dining room and a kitchen plus, of course, a bath, bathroom and entryway) squeezed into, perhaps, 700 square feet. But the price was so high that banks were beginning to write 60-year '3-generation' mortgages which would be finally paid off by the grandson of the signer. With prices at such a level, inheritance was a problem. The tax on inheritances ran to 70% at the top end and, with many homes valued toward a million dollars, most salarymen were forced into deep debt or would have to sell the family home to pay the tax.
In the early 1970s, pollution was a problem, especially in large cities like Tokyo. Policemen were provided with oxygen if they had to direct traffic at a busy intersection and smog alerts rose from 16 a year in 1960 to 150 a year in 1970: school children and those with breathing problems were advised not to go outside on those days. To a large extent, air pollution has been solved and from time to time Tokyo's residents can even see Mt. Fuji once again.
What has not been solved is commuting. Even if the commuters' ribs are not broken, the time spent on public transport has increased from about 60 minutes in 1960 to and average of 90 minutes. The time spent is a serious degradation of the quality of life, even if the trains run exactly on time, are air conditioned, and have stock quotations and weather forecasts on LCD monitors at each door.
As the economy has grown, pressures have increased to deal with questions of living conditions and the quality of life. Pollution has been reduced and more households are connected to the sewer system in the cities, but there is pressure on local and national governments to invest more to improve living conditions. Local governments especially have been responsive, but the national government controls the bulk of public money and therefore long-term improvement depends on it.
The unofficial
It's not really possible in the US. I recently saw a documentary about the progress in the robotics and it contained one explanation why quasi-androids are being so expensivelly developed in Japan and not in the US.
Basically Japan is a closed country with its population getting older every day which makes the workforce very expensive there today and even more expensive in future.
US on the other hand is still a country open for immigrants with hordes of young people from all over the world willing to work for food. Or even cheaper. And if it's still too expensive US outsources the work to the third world countries.
There's no place for robots in US economy.
rrw
Bastard Operator From 193.219.28.162
Seriously, thought, what's with all this madness of making robot as close to a human as possible? Humans as such are quite imperfect and while it is quite a nice to impress people around, it's practicall applications are rather fare fetched.
Building robot that can intelligently mowe your lawn (without need for special costly installation), deliver pizza in a building etc.and dynamically react to it's enviroment is much more usefull and almost equally as hard.
These Japan guys should consider if they want to put their money into impressing the world or making usefull, althought not so ipmressive, technology.
"Two beers or not two beers. That's the question." -- Shakesbeer
I remember the excitement in the U.S. AI community when Feigenbaum went to Japan and sold the government there on the 5th generation build-a-real-AI project.
Funny - I do not remember any animosity - mostly just wishing them good luck.
BTW, the 5th generation project was built around logic programming (Prolog variants). I have never understood why more people do not use Prolog. For an admittedly small percentage of software projects, Prolog is the best language for solving problems - well worth learning. (A very good free LGPL Prolog is available here).
-Mark
In one episode of The Simpsons, Homer decides to set up a bar in his garage. Dialog paraphrased:
Scene: Homer is clearing out the garage.
Bart: "Is this one of those projects you start and never finish?"
Homer: "Hey, when I start something, I stick with it to the end!"
Homer removes a box, revealing a pathetic robot with a bucket body and mismatched arms, one made from a broom. It looks exactly like what Homer would come up with if he decided he wanted a robot boy.
Robot: "FA-THER! GIVE ME LEGS!"
Homer: "I thought I told you to clear out!"
He grabs the robot and tosses it into the road.
Robot, trailing modules from his open lower torso, drags himself away. He pauses and looks back, but Homer points firmly down the street.
((Shudder))
* * *
What's great about this vignette: It could have been done in 1964, by "The Other Limits" or "The Twilight Zone," only they would need a full hour.
In the hands of Groening and company, this drama of horrifying pathos gets boiled down into a throw-away segment lasts thirty seconds, tops.
* * *
And, um, to make this topical: Given the Japanese tendency toward faddishness, I fully expect the garbage dumps of Tokyo thirty years down the line to be swarming with last year's model of robot child.
(I actually wrote a story about something similar; American kid discovers that the neighborhood lawn-care robots are repurposed My Buddy Dragon and My Pretty Lioness playmate 'bots, shorn of their cosmetic foam rubber shells and sealed in utilitarian green plastic skins.)
One can "attempt" something and still be successful. The submitter is comparing the Japanese efforts to American efforts, suggesting that good may come from the efforts, if not the final product.
It's nothing but crumpled porno and Ayn Rand.
50 billion is not "a quarter billion $US", unless you use your own definition of a quarter. It's actually more like 423 million $US or 381 million Euro*.
* - Sorry, fucked up Slashcode doesn't support "advanced" non-ASCII characters, like a euro symbol.
Future Wiki -- If you don't think about the future, you cannot have one.
Is that it assumes that with constant progression you can reach a given goal. They set the goal seemingly-low -- "A five year old" -- to make it seem more achievable, when the real problem is that to achieve this it's going to take a paradigm shift in technology. This shift could occur thirty years from now, a hundred years from now, or tomorrow, but it's not something you can put on a schedule.
The idea of setting a "five-year-old" requirement on it is ridiculous, because what we lack is the basic ability to create human-like intelligence in the first place. Once we have that, it will be trivial to make the equivalent of a five year old human (basic sentience), or a fifteen year old human (the peak of human intelligence), or even something beyond that that humans are incapable of achieving (After fifteen years of age or so human intelligence goes into a slow downward slide, though overall capability often goes up thanks to accumulated experience and knowledge. Imagine a being that had equal or greater intelligence to a fifteen year old, but with the knowledge and life experience of a fifty year old!).
It's kind of like the development of the microprocessor. Before we knew how to make one, there was nothing -- but once we had the basic technology to make one, Moore's law kicked in and the capability of microprocessors grew by leaps and bounds. AI will be the same way. Once we have a big breakthrough that allows us to create the first real AI, the technology will progress with incredible rapidity. The problem is that first big breakthrough, and it's not something you can simply budget time and resources for and expect results. You can't put it on a "thirty year plan".
Cool, they are making 5-year-olds obsolete.
You presume that a nearly infinite amount of power is available. While it is conceivable that fusion power may one day become feasable, it still seems very far away indeed. Until the power problem is solved, there really isn't a chance that zillions of robots could be produced and deployed doing all sorts of manual labor.
Evolution created me out a pool of chemicals.
Eh, I just checked, and 50 billion Yen equal $US 423 million, not $US 400 billion.
Wow, $US400 billion every year, that would be more than 10% of their total purchasing power (quoting CIA's numbers), and about 90% of their total gross revenue (not yet calculating their expenditure). That would have been some serious fucking spending. But no, they're not spending that many dollars, it's just the story submitter's inability to do math.
What time is it/will be over there? Check with my iPhone app!
Perhaps the U.S. Government should consider funding such a program over here?
5 year old minded robot? We already have Bush.
The current (8/20/2003) exchange rate is 118.015 yen to the dollar.
Conceptual Guerilla is a great site with a decidedly leftist political bent that attempts to expose and digest some of the consequences of this new reality. I suggest anyone who's interested in discussing this further to head over to the forums there. I'd also like to thank the Slashdotter who put this link in their sig.
"I assumed blithely that there were no elves out there in the darkness"
I'm not sure how Japan figures out what to pick but it seems to work. Maybe they are making very good choices or maybe if you stick enough money into something it will eventually pay off. And as sceptical I am of humanoid robots I can't say this is a silly idea any more.
One Word: CONSTRUCTION
Although my own insticts in this regard go to smaller construction-bots (say mouse or even cockroach sized), using loosely coupled swarming network communication and behavior to build. Probably cheaper and more powerful. But, let em try... Even if it's a stupid idea, if no one else competes, they will win.
The reason that it can be true that 1+1 > 2 is that very peculiar nonzero value of the + operator
Not to rub on a fine point but a Yen is about a penny. US$400B would require about Y47 TRILLION on today's market. The point the article is attempting to make: that Japan is making an equivalent investment as the USA did with Apollo. That implies percent of GDP. 1960's Japan cannot remotely be compared to 1960's USA in terms of economic capacity, so more than exchange rate must be used. How about adjusting that 1969 exchange rate in terms of real GDP? Hell, in 1965 the US economy was about $3T, making the $19B budget for Apollo about .6% of GDP. That same portion in 2003 terms would be about $70B. Apollo lasted 9 years averaging about $2.2B per year, or, in 2003 terms, roughly $7B per year, or about Y826B per year. This program at Y50B per year over 30 years would be about Y1.5T total. If compacted down from 30 to 9 years, like Apollo, this would still only be Y450B -- about half. The $12.7B this project will cost in 2003 dollars would be about $3.5B in 1969, or ~.1% of GDP in either year -- one-sixth the investment in Apollo in terms of percent of GDP and still just over half in terms of total dollars unadjusted for over 30 years of inflation. Still f'ing huge, but nowhere near Apollo.
Our children should do nothing for a living.
There isn't any reason to concoct something for them to do.
They should simply be educated on the dangers of over-population and the use of contraceptives and how to operate the robots.
That's it.
"I assumed blithely that there were no elves out there in the darkness"
Yes, then forget about the robots and colonize the moon or Mars.
Though lets start with a REAL space station first.
Actually, lets start with a more dependable heavy payload launch vehicle.
Three seperate posts, saying in effect, "Who cares what we spend the money on, as long as the government spends, spends, SPENDS!"
I, personally, agree with the spirit of the first poster who reccomends that we "worry about the robots after we figure out how to pay back our debt." (Although, it does look like Krisp wants to spend money on state-sponsored "education" - you have to have gone through a US public school to appreciate the irony in that.) And that's currently modded funny?
It's my money. Is it so wrong to let me keep it?
Carthago delenda est!
...to get robots with the capabilities of five year olds. Just clone Congress a couple of times.
-- "For every complex problem, there is a solution that is simple, neat and wrong." -- HL Mencken