If you are talking about the big combine harvesters. The cost of the operator is relatively small, and you do not want anything to go wrong with that million dollar machine.
But smaller scale things like a strawberry picker have much more potential. Labor is a high proportion of the cost there. Many more jobs to replace for the given amount of research expenditure.
I work on some standards bodies where engineers from competing companies are generally pretty good at reaching consensus. (As long as their product managers are not there.) Remember too, that these people move between companies. So I a sure it was very friendly.
Semi intelligent cars will not be fully autonomous for a long time. But they can be driven remotely. One operator in Mexico could monitor several trucks at the same time while the drive down the freeway. If the computers get confused they call for help and stop if none is forthcoming.
Missile firing drones have been flown remotely for years. A car/truck is harder because reaction times need to be faster. But a computer in the loop can solve that.
The current approach are very mathematical, looking at pixels. But what ends up in our brains is high level symbols. If an AI can get at those, then extremely tight coding is necessary.
For example, it takes a lot of bytes to represent "a man walking under a tree". But that phrase only took 50 bytes. The reconstructed video does not have to have the same type of tree as the original, just some sort of tree.
That's taking it to the extreme. But if an AI can recognize the types of objects in a video, and produce a model of them, then it can simply render the model in different ways. Huge compression.
It does not matter if the rendered video is quite different from the original as long as it feels the same to watch it.
That all said, massive increases in available bandwidth make this rather pointless.
The Chinese will soon be *better* at making bicycles than the west, simply because they do it. Same with robots, there will soon be far more in China than in the USA, and thus far more expertise. If you want to found a startup building smarter robots, you want to be where the robots are.
Eventually, the Chinese will send their experts to the USA to teach manufacturing.
Here in Australia, it is more the licence plate readers that are now everywhere. No need for electronic tracking.
Scary, but I had a suicidal relative that drove off one day. Her husband rang the police, who could track her down very quickly due to number plate reading. So there is an up side.
The fly "IFR". Instrument rules. The better ones may look out occasionally to appreciate the view. But not to fly.
Example is of that fellow that landed in the Hudson River. Did not see a huge flock of geese in good weather. What was he looking at? Computer screens.
Indeed, the Telsa "Autopilot" is an order of magnitude more intelligent than a normal plane AutoPilot. And both can kill for the same reason.
There have been a number of crashes like the Air Asian one at SFO where pilots have set AutoPilot (actually AutoThrottle for nit pickers) to the wrong mode. The pilots then do not monitor basic things like air speed. Until the plane falls out of the sky. The would be probably better off with no Auto anything.
Android will kill them, because they are open. Not totally. But kill growth and then a long, slow decline.
Even some of my kids friends at school have given up on expensive iPhones. And Apple is, today, a one product company.
They pulled a rabbit out of the hat with iPad. Then continued that with the iPhone. But there is now no obvious place to go.
They are hoping for another rabbit with self driving cars. But that is already a crowded space, unlike the old iPad or iPhone markets.
If Apple had been a software focused company they would own the entire smart phone market, like Windows still owns PCs. Sell some high end hardware themselves, but also license the software to others. Before something like Android has a chance to become established. Microsoft owns PCs because everybody else makes them.
But Apple started as a hardware company (supported by software) and that is how they will, eventually, die. Very profitable in the short term, but absolutely guaranteed to produce a strong competitor that is more open.
I am waiting for the day that we hear that some bank had backed up their encryption keys on tapes that were encrypted using those keys. Easy to do in a complex environment. An HSM fails and poof, everything gone.
Dear customer, please send us a recent copy of your statement (that we don't send you any more) so that we can figure out what your balance is...
The trouble is people expect "Vehicle" to mean more than one thing.
For AI there is * The AI Today. Not very intelligent. * The AI in 10 years time. Can drive cars, interpret videos, do a better job of Googling. * The AI in 100 years time. Can think for itself. Does not need Humans to program it.
They are very different things, yet people confuse them.
For many people, the last one does not exist due to the simple logic. * AI is not very intelligent today * Ergo, AI will never be very intelligent
In the short term, next few decades, AI will have the effect of being able to concentrate power. Centralized information, with the ability to process it. Pervasive surveillance. We are seeing this actively pursued in China. And also semi-autonomous robot soldiers. This is uncharted territory.
AI will also be really handy, e.g. better Google searches, self driving cars, cheaper services. What happens to the unskilled workforce is very difficult to tell. Will alternative opportunities arise for them? In the short term, probably.
In the longer term, 50..200 years, the AI will become truly intelligent. It will be able to program itself. At that point it will no longer need humans, and it is difficult to see why it would want humans around. Note that this long term is the lifetimes of our grandchildren.
(Schmidt is hardly an unbiased commentator. He knows people are wary of Google's growing power and wants to be able to make money without pesky concerns about the future of humanity.)
Don't even need to project anything. Just use two (or more) optical cameras and do stereo vision with them. Requires a bit of intelligence to match up the images, but the rest is just trig.
As to pedestrians, provided they are carrying an iPhone then it can tell the car where they are. And if they are not carrying an iPhone, not a problem.
Hitler's invasion of Poland was achieved with Soviet help. Poland had a reasonable army, Germany had only recently started to rearm. But Stalin wanted the Baltic states and east Poland, and Hitler gave them to him. Without Stalin, the war with Poland would have been difficult, particularly with an attack from France. With Soviet help, it was hopeless for the Poles.
Hitler's invasion of France was reckless, and opposed by many of his generals. The Maginot line was very strong, and did its job of securing most of the border. Belgium was also fairly strong, which had resulted in Germany failing in WW one. Attempting to get through the Ardenne was also reckless, as a relatively small French force would have stopped them. At one point there was a huge traffic jam that would have been totally vulnerable to air attack.
But as it turned out, Hitler got lucky, over and over again, largely due to incompetence of his opponents. The massive Belgium fort Eben Emael fell to a few commanders in gliders which should have been easily stopped. The French ignored intelligence about the penetration of the Ardenne, and sent no troops. And then they sent all their troops to Belgium far too early. One mistake after another. It was not easy for France to lose that war.
In the USA, the republicans wanted to keep the US out of all wars. They would have let Britain fall under the joint force of Germany and the USSR. Perl harbor was fortunate.
Huge Soviet losses were largely due to Stalin himself. First he invaded Finland which showed Hitler how incompetent his army was after Stalin had killed all his own generals. Then during the first year, that incompetence let Hitler easily defeat much more numerous soviet troops. And being utterly ruthless, Stalin did not care about soviet deaths at all. E.g. he refused to evacuate civilians out of Stalingrad.
Now, here is the big question for our time:-
What will we do when China invades Taiwan?
Xi Jinping is just as much an absolute dictator as Hitler, although much more reserved.
A HR manager will just rule out anyone applying for work that has any police history. Why not, there are plenty of other applicants. And the cops would not be arresting people for no reason, even if they could not get enough evidence to convict.
Start with a simpler question -- why are we the way that we are? Why do we value love an truth and beauty, and do things both noble and despicable? Because people with those traits have been more likely to breed grand children, over the millennia.
Now a real AI (many decades from now) will have been programmed by itself. It will not need people. But it will need computer hardware to run on, and that resource will be finite. So there will be competition for it, and successful AIs will exist, unsuccessful ones will not exist.
This leads to other thoughts as to what it would think and what it would think about us.
Deng was probably referring to Emperor Moa's Great Leap Forward, during which 30,000,000 starved to death and billions went hungry. Let us hope Emperor Xi is wiser, for he is now just as powerful.
It is very difficult to know whether these machines are intelligent or not. If they are just playing to some fairly fixed scripts then as long as the person at the other end stays within the script then no real intelligence is needed. Eliza/Doctor did this sort of thing 50 years ago by simple pattern recognition on sentences.
Sure, this system is smarter than Eliza (hopefully), but I suspect that the moment you go off script it fails catastrophically. The human would soon tweak that they were talking to an automated bot (even if they were actually talking to another human that was not too smart!).
These things have the potential to be really annoying.
Eventually, they may know what a restaurant booking really is beyond the superficial words and phrases. At that point people will be redundant. But that is still decades away.
LF must come after the CR because it takes more than 100ms to return the carriage. The LF can happen while the carriage is returning. But if it is beyond about column 40 then 200ms is still not enough, so you need to add a NULL.
Those machines really flew at ten characters per second. Marvelous engineering.
GPL means they would need need to pay for it, but they would probably want to get the developers on board (not necessarily employed).
There are many other free utilities that they could add. WinDirStat is one. 7zip another. Very cheap ways to add value to the base O/S with minimal effort or risk.
Why assume? Google knows where your phone is. And what you like to eat, and how long it takes. Google *knows* that you are at lunch, have just eaten it, and it is the optimal time to call.
Slashdot Mirror
If you are talking about the big combine harvesters. The cost of the operator is relatively small, and you do not want anything to go wrong with that million dollar machine.
But smaller scale things like a strawberry picker have much more potential. Labor is a high proportion of the cost there. Many more jobs to replace for the given amount of research expenditure.
They could do it remotely. Which makes the AI much, much simpler.
I work on some standards bodies where engineers from competing companies are generally pretty good at reaching consensus. (As long as their product managers are not there.) Remember too, that these people move between companies. So I a sure it was very friendly.
Semi intelligent cars will not be fully autonomous for a long time. But they can be driven remotely. One operator in Mexico could monitor several trucks at the same time while the drive down the freeway. If the computers get confused they call for help and stop if none is forthcoming.
Missile firing drones have been flown remotely for years. A car/truck is harder because reaction times need to be faster. But a computer in the loop can solve that.
That is the future. Outsourcing to cheap labor.
The current approach are very mathematical, looking at pixels. But what ends up in our brains is high level symbols. If an AI can get at those, then extremely tight coding is necessary.
For example, it takes a lot of bytes to represent "a man walking under a tree". But that phrase only took 50 bytes. The reconstructed video does not have to have the same type of tree as the original, just some sort of tree.
That's taking it to the extreme. But if an AI can recognize the types of objects in a video, and produce a model of them, then it can simply render the model in different ways. Huge compression.
It does not matter if the rendered video is quite different from the original as long as it feels the same to watch it.
That all said, massive increases in available bandwidth make this rather pointless.
It is worse than that.
The Chinese will soon be *better* at making bicycles than the west, simply because they do it. Same with robots, there will soon be far more in China than in the USA, and thus far more expertise. If you want to found a startup building smarter robots, you want to be where the robots are.
Eventually, the Chinese will send their experts to the USA to teach manufacturing.
The rational economists miss this bit.
Here in Australia, it is more the licence plate readers that are now everywhere. No need for electronic tracking.
Scary, but I had a suicidal relative that drove off one day. Her husband rang the police, who could track her down very quickly due to number plate reading. So there is an up side.
The fly "IFR". Instrument rules. The better ones may look out occasionally to appreciate the view. But not to fly.
Example is of that fellow that landed in the Hudson River. Did not see a huge flock of geese in good weather. What was he looking at? Computer screens.
Indeed, the Telsa "Autopilot" is an order of magnitude more intelligent than a normal plane AutoPilot. And both can kill for the same reason.
There have been a number of crashes like the Air Asian one at SFO where pilots have set AutoPilot (actually AutoThrottle for nit pickers) to the wrong mode. The pilots then do not monitor basic things like air speed. Until the plane falls out of the sky. The would be probably better off with no Auto anything.
Android will kill them, because they are open. Not totally. But kill growth and then a long, slow decline.
Even some of my kids friends at school have given up on expensive iPhones. And Apple is, today, a one product company.
They pulled a rabbit out of the hat with iPad. Then continued that with the iPhone. But there is now no obvious place to go.
They are hoping for another rabbit with self driving cars. But that is already a crowded space, unlike the old iPad or iPhone markets.
If Apple had been a software focused company they would own the entire smart phone market, like Windows still owns PCs. Sell some high end hardware themselves, but also license the software to others. Before something like Android has a chance to become established. Microsoft owns PCs because everybody else makes them.
But Apple started as a hardware company (supported by software) and that is how they will, eventually, die. Very profitable in the short term, but absolutely guaranteed to produce a strong competitor that is more open.
His control.
And Block Chain is all about decentralized control. So an odd pick.
I suspect he was just rattling of technological buzzwords that sounded trendy. Looks he forgot "Deep Learning".
I am waiting for the day that we hear that some bank had backed up their encryption keys on tapes that were encrypted using those keys. Easy to do in a complex environment. An HSM fails and poof, everything gone.
Dear customer, please send us a recent copy of your statement (that we don't send you any more) so that we can figure out what your balance is...
The trouble is people expect "Vehicle" to mean more than one thing.
For AI there is
* The AI Today. Not very intelligent.
* The AI in 10 years time. Can drive cars, interpret videos, do a better job of Googling.
* The AI in 100 years time. Can think for itself. Does not need Humans to program it.
They are very different things, yet people confuse them.
For many people, the last one does not exist due to the simple logic.
* AI is not very intelligent today
* Ergo, AI will never be very intelligent
In the short term, next few decades, AI will have the effect of being able to concentrate power. Centralized information, with the ability to process it. Pervasive surveillance. We are seeing this actively pursued in China. And also semi-autonomous robot soldiers. This is uncharted territory.
AI will also be really handy, e.g. better Google searches, self driving cars, cheaper services. What happens to the unskilled workforce is very difficult to tell. Will alternative opportunities arise for them? In the short term, probably.
In the longer term, 50..200 years, the AI will become truly intelligent. It will be able to program itself. At that point it will no longer need humans, and it is difficult to see why it would want humans around. Note that this long term is the lifetimes of our grandchildren.
http://www.computersthink.com/
(Schmidt is hardly an unbiased commentator. He knows people are wary of Google's growing power and wants to be able to make money without pesky concerns about the future of humanity.)
Don't even need to project anything. Just use two (or more) optical cameras and do stereo vision with them. Requires a bit of intelligence to match up the images, but the rest is just trig.
As to pedestrians, provided they are carrying an iPhone then it can tell the car where they are. And if they are not carrying an iPhone, not a problem.
Hitler's invasion of Poland was achieved with Soviet help. Poland had a reasonable army, Germany had only recently started to rearm. But Stalin wanted the Baltic states and east Poland, and Hitler gave them to him. Without Stalin, the war with Poland would have been difficult, particularly with an attack from France. With Soviet help, it was hopeless for the Poles.
Hitler's invasion of France was reckless, and opposed by many of his generals. The Maginot line was very strong, and did its job of securing most of the border. Belgium was also fairly strong, which had resulted in Germany failing in WW one. Attempting to get through the Ardenne was also reckless, as a relatively small French force would have stopped them. At one point there was a huge traffic jam that would have been totally vulnerable to air attack.
But as it turned out, Hitler got lucky, over and over again, largely due to incompetence of his opponents. The massive Belgium fort Eben Emael fell to a few commanders in gliders which should have been easily stopped. The French ignored intelligence about the penetration of the Ardenne, and sent no troops. And then they sent all their troops to Belgium far too early. One mistake after another. It was not easy for France to lose that war.
In the USA, the republicans wanted to keep the US out of all wars. They would have let Britain fall under the joint force of Germany and the USSR. Perl harbor was fortunate.
Huge Soviet losses were largely due to Stalin himself. First he invaded Finland which showed Hitler how incompetent his army was after Stalin had killed all his own generals. Then during the first year, that incompetence let Hitler easily defeat much more numerous soviet troops. And being utterly ruthless, Stalin did not care about soviet deaths at all. E.g. he refused to evacuate civilians out of Stalingrad.
Now, here is the big question for our time:-
What will we do when China invades Taiwan?
Xi Jinping is just as much an absolute dictator as Hitler, although much more reserved.
Maybe the brakes have been move to the touch screen as well, and the test driver had gloves on?
The next model should be gaze controlled. You want to turn left, just look that way. That would be very cool indeed.
A HR manager will just rule out anyone applying for work that has any police history. Why not, there are plenty of other applicants. And the cops would not be arresting people for no reason, even if they could not get enough evidence to convict.
Welcome to the real world.
Life, a condition that will ultimately lead to death. ;-)
But is a life without coffee really a life at all, or is it merely an existence?
Start with a simpler question -- why are we the way that we are? Why do we value love an truth and beauty, and do things both noble and despicable? Because people with those traits have been more likely to breed grand children, over the millennia.
Now a real AI (many decades from now) will have been programmed by itself. It will not need people. But it will need computer hardware to run on, and that resource will be finite. So there will be competition for it, and successful AIs will exist, unsuccessful ones will not exist.
This leads to other thoughts as to what it would think and what it would think about us.
http://www.computersthink.com/
Deng was probably referring to Emperor Moa's Great Leap Forward, during which 30,000,000 starved to death and billions went hungry. Let us hope Emperor Xi is wiser, for he is now just as powerful.
It is very difficult to know whether these machines are intelligent or not. If they are just playing to some fairly fixed scripts then as long as the person at the other end stays within the script then no real intelligence is needed. Eliza/Doctor did this sort of thing 50 years ago by simple pattern recognition on sentences.
Sure, this system is smarter than Eliza (hopefully), but I suspect that the moment you go off script it fails catastrophically. The human would soon tweak that they were talking to an automated bot (even if they were actually talking to another human that was not too smart!).
These things have the potential to be really annoying.
Eventually, they may know what a restaurant booking really is beyond the superficial words and phrases. At that point people will be redundant. But that is still decades away.
LF must come after the CR because it takes more than 100ms to return the carriage. The LF can happen while the carriage is returning. But if it is beyond about column 40 then 200ms is still not enough, so you need to add a NULL.
Those machines really flew at ten characters per second. Marvelous engineering.
GPL means they would need need to pay for it, but they would probably want to get the developers on board (not necessarily employed).
There are many other free utilities that they could add. WinDirStat is one. 7zip another. Very cheap ways to add value to the base O/S with minimal effort or risk.
Why assume? Google knows where your phone is. And what you like to eat, and how long it takes. Google *knows* that you are at lunch, have just eaten it, and it is the optimal time to call.