Supersmart Robots Will Outnumber Humans Within 30 Years, Says SoftBank CEO (fortune.com)

Computers running artificial intelligence programs will exceed human intelligence within three decades, Masayoshi Son, founder of the Japanese technology and telecommunications conglomerate SoftBank Group, said on Monday. From a report on Fortune: "I really believe this," Son told a large audience at the Mobile World Congress, the telecom industry's annual conference in Barcelona. A computer will have the IQ equal to 1,000 times the average human by that point, he said. Even clothing like a pair of sneakers will have more computing power that a person, Son joked. "We will be less than our shoes," he said, to laughter. Asked if the rise of the computer could be dangerous for humankind, Son said that would be up to how people react. "I believe this artificial intelligence is going to be our partner," he said. "If we misuse it, it will be a risk. If we use it right, it can be our partner."

Comments

Maybe they will be smart enough

[Chrisq]

Maybe they will be smart enough to be able to tell the difference between "outnumber" and "outsmart", thus outsmarting the poster

Re:Maybe they will be smart enough

[Anonymous Coward]

Especially criminal since the poster was trying to reword the perfectly decent headline from the linked article, "Computers Will Be Smarter Than People In 30 Years". Why not just quote the headline, since you're linking to the article, instead of butchering it?

Re:Maybe they will be smart enough

[Anonymous Coward]

... or the slashdot editor I should say, in case the headline butchery wasn't the work of the person who submitted the article.

Re:Maybe they will be smart enough

[Notabadguy]

Maybe they will be smart enough to be able to tell the difference between "outnumber" and "outsmart", thus outsmarting the poster

Hopefully also smart enough to proofread their own sentences, so we don't end up with garbage like "more computing power that a person."

Re: Maybe they will be smart enough

[Notabadguy]

Slashdot editors: "I click a button, they send me a paycheck."

As a friend of mine said long ago referring to a construction worker holding up a sign that said "Slow" (as in "slow down your car"): you could be replaced by a bucket of sand.

Your friend is forgetting the other half of the job....which is the ultra-complex maneuver of coordinating with another sign holder and rotating the sign from "Slow" to "Stop." That immensely complex operations requires qualifications unlikely to be met outside of highly paid human labor for generations to come.

Smart enough to REALLY f*ck things up???

[BarbaraHudson]

IQ is not a real measure of intelligence. Witness the fact that 99.7% of geniuses are not stupid enough to pay MENSA $60 a year for a card saying "I'm smart." That 99.7% realize that it's possible to be really intelligent AND do really dumb things at the same time - they just have to look at MENSA members.

And the last time I pointed this out, along came all the MENSA members saying how it isn't so. Proving that Dunning-Kruger is no respecter of IQ tests. :-)

Now if they could create devices that showed more common sense than, say, Donald Trump (I know, I set the bar REALLY low, but you've got to start somewhere) they might have something.

Re:Smart enough to REALLY f*ck things up???

[m00sh]

IQ is not a real measure of intelligence. Witness the fact that 99.7% of geniuses are not stupid enough to pay MENSA $60 a year for a card saying "I'm smart." That 99.7% realize that it's possible to be really intelligent AND do really dumb things at the same time - they just have to look at MENSA members.

And the last time I pointed this out, along came all the MENSA members saying how it isn't so. Proving that Dunning-Kruger is no respecter of IQ tests. :-)

Now if they could create devices that showed more common sense than, say, Donald Trump (I know, I set the bar REALLY low, but you've got to start somewhere) they might have something.

IQ is like height in basketball. The best basketball players aren't the tallest people in the world but they are all taller than average.

There is a certain height above which isn't advantageous in basketball. Same with IQ. There is a good enough IQ and beyond that doesn't matter.

Also, two people with high IQ will out-perform a single person of super high-IQ. If a team with less skilled basketball players is allowed to play with an extra player, they will beat a team with better players. So, the social environment that allows people to work together is more important than finding people of super-high IQs.

Also, there have been lots of data collected on IQs and success. The highest correlation to success wasn't IQ, it was how successful the parents were. If you parents can provide you a good learning environment and access to connections, it is more important than just being smart.

Re: Smart enough to REALLY f*ck things up???

[Anonymous Coward]

No offence barbarahudson but u should have your IQ looked at. I predict you're significantly less smary than you think. IQ isnt the full story, you're right about that. It might one day surprise you- what else you're missing.

Re: Smart enough to REALLY f*ck things up???

[Anonymous Coward]

Someone's jelly.

Re: Smart enough to REALLY f*ck things up???

[Anonymous Coward]

smart* even. Baited

Re:Smart enough to REALLY f*ck things up???

[reboot246]

I don't know what my IQ is, but I'm smart enough to know that I don't need a pair of sneakers with AI.

Just because we can doesn't mean we should.

Re:Smart enough to REALLY f*ck things up???

[ShanghaiBill]

they just have to look at MENSA members ...

It sounds like you have some serious insecurity issues.

Re:Smart enough to REALLY f*ck things up???

[BarbaraHudson]

What are you going to do when those super-smart sneakers get hacked and decide you should go somewhere else? Or the battery catches fire and they won't untie themselves? Or you become the target of thieves because if you can afford to waste money on intelligent sneakers (because you obviusly have more money than brains) what other swag can they snag off you? Or if someone SWATs your sneakers? You're walking along and your sneakers are broadcasting alerts to everyone in the area that you're a wanted child molester?

On second thought, I can't wait until our idiot overlords start going around with their intelligent sneakers.

Re:Smart enough to REALLY f*ck things up???

[BarbaraHudson]

Nah, I'm in the 99.7% who don't need a MENSA card. You, on the other hand ... see my point about Dunning-Kruger. On second thought, don't bother - it's like the joke about the roof - it's over your head. :-)

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

Just in case you don't know, Mensa caters to high IQ people who haven't accomplished anything. It's a way to buy self esteem, similar to the way gun nuts buy theirs. I belonged when I was young and hadn't accomplished anything yet. It took me quite a few years of hard work build a life I could actually be proud of. Until then Mensa was a good substitute and cheaper than guns.
     

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

You're walking along and your sneakers are broadcasting alerts to everyone in the area that you're a wanted child molester

1989, that number! Another summer! Sound of the funky drummer! Fight the power!

Re:Smart enough to REALLY f*ck things up???

[Bengie]

two people with high IQ will out-perform a single person of super high-IQ

Out-perform them in what? One of the biggest issues in programming is a group of programmers is only as smart as the least smart of the group, unless the rest of them just ignore the lesser. I've had it where about 5 of the most senior programmers in my company where stuck a problem for several days with code that they wrong. I just so happened to have walked by and overheard them discussing the issue and I solved the problem in less than a minute. It was a 100% custom built system, so I had zero knowledge or experience. This pretty much sums up every experience I've ever had.

Another example. My ISP was having a network issue. I was working with one of their senior engineers for nearly 6 months. I kept telling them what the problem was and where they need to look, but they couldn't figure it out. They eventually hired a 3rd-party firm that specialized in fixing complex network issues for large ISPs. Even they could not figure it out for several days. My ISP called me and asked me to talk to these people. I got on the phone with 3 "specialists", and I went over the data I sent them, explaining what I thought the issue was and the reasoning behind my thoughts. I got off the phone, 1 hour later, the problem was figured out.

My background in networking is I took an entry level class in college and I have a pfSense box at home, and I diagnosed an ISP level network issue with nearly zero experience.

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

IQ is not a real measure of intelligence. Witness the fact that 99.7% of geniuses are not stupid enough to pay MENSA $60 a year for a card saying "I'm smart." That 99.7% realize that it's possible to be really intelligent AND do really dumb things at the same time - they just have to look at MENSA members.

And the last time I pointed this out, along came all the MENSA members saying how it isn't so. Proving that Dunning-Kruger is no respecter of IQ tests. :-)

Now if they could create devices that showed more common sense than, say, Donald Trump (I know, I set the bar REALLY low, but you've got to start somewhere) they might have something.

IQ is like height in basketball. The best basketball players aren't the tallest people in the world but they are all taller than average.

There is a certain height above which isn't advantageous in basketball. Same with IQ. There is a good enough IQ and beyond that doesn't matter.

Also, two people with high IQ will out-perform a single person of super high-IQ. If a team with less skilled basketball players is allowed to play with an extra player, they will beat a team with better players. So, the social environment that allows people to work together is more important than finding people of super-high IQs.

Also, there have been lots of data collected on IQs and success. The highest correlation to success wasn't IQ, it was how successful the parents were. If you parents can provide you a good learning environment and access to connections, it is more important than just being smart.

You bring up an important point in this discussion, this premise that computers will out perform humans in this thing called intelligence is dependent on how you define intelligence. Currently intelligence is not well defined, which is why building artificial intelligence has had such a slow return on investment for all the effort that has been put into it. IQ is an oversimplification of metrics intended to measure general brain power and even that is a loose definition of something that they cannot define in concrete terms. Until we have a general definition for intelligence and a metric for accurate and concise measurement of it objectively it will always be a woo woo thing that remains in the realm of science fiction.

We have made progress on building machines to do tasks that seem intelligent, such as playing Chess or GO or some other game, but can those programs do your taxes or paint a compelling picture or write a piece of music? There is a big difference between a machine being proficient in a narrow task and having general "intelligence" is my point. You can also make the point that even though the hardware is up to the task, if such a super computer is not intelligently programmed or trained, it will not appear intelligent, whatever that may mean at the time.

Re:Smart enough to REALLY f*ck things up???

[pr0fessor]

My brother joined mensa when he was in college so he could mess with them.... of course he is brilliant and mad as a hatter.

Re: Smart enough to REALLY f*ck things up???

[Anonymous Coward]

Did you turn out to be right? It's pretty important to determine whether that is a "cool story bro" or "you need to charge more for your brilliance."

Re:Smart enough to REALLY f*ck things up???

[HiThere]

IQ is, indeed, not a good measure of intelligence. In fact, intelligence isn't a unitary thing, but a bunch of separate capabilities, at least one of which handles organizing and communicating with the other parts.

That said, if we're going to talk informally about intelligence, IQ is a reasonable stand-in. It means something pretty reasonable in the area between 80-120, possibly 75-125. I'll grant that in no area is is a really good definition, but it's easily quantified.

Note that the very concept of an IQ of 1000 doesn't make any sense. So accept it as a figure of speech. Accepting it as a figure of speech, I still think he's wrong, because I believe that for every task there is an optimum level of intelligence. If he's approximately correct, then there will be a very few extremely intelligent AIs, but it sure won't be your sneakers. The claim that it *could* be in my sneakers is interesting, and a bit unbelievable. And I've got large feet. (Well, he didn't claim that the super-AI would be in my sneakers, just that it would have more computing power than I did, which is also a bit unbelievable unless you start doing strange things with word definitions. I could manage definitions that would make that a reasonable claim, but they sure aren't the standard ones.)

OTOH, my projection to a human equivalent AI is still around 2030, which is sooner than he is talking about. But I'm not expecting that thing to be mobile or portable. And when I say "human equivalent" I'm not talking about all characteristics. I'm not talking about motivational structure. I'm not talking about built-in sense organs. I'm not talking about computations/watt. Etc. I'm talking mainly about ability to reason about situations with incomplete data of uncertain reliability...which, admittedly, covers a lot of what we do.

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

IQ is not a real measure of intelligence. Witness the fact that 99.7% of geniuses are not stupid enough to pay MENSA $60 a year for a card saying "I'm smart." That 99.7% realize that it's possible to be really intelligent AND do really dumb things at the same time - they just have to look at MENSA members.

And the last time I pointed this out, along came all the MENSA members saying how it isn't so. Proving that Dunning-Kruger is no respecter of IQ tests. :-)

My father was a MENSA member in the '70s. Apparently they had a lot of parties back then with strippers. That's why he went.

Now if they could create devices that showed more common sense than, say, Donald Trump (I know, I set the bar REALLY low, but you've got to start somewhere) they might have something.

Donald Trump is rich and President of the United States, and you're not. Go figure.

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

I think a 12' tall basketball player might be good...if his joints can handle the weight...

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

they just have to look at MENSA members ...

Why bring sexual relations with minors into this?

Re:Smart enough to REALLY f*ck things up???

[BarbaraHudson]

My father was a MENSA member in the '70s. Apparently they had a lot of parties back then with strippers. That's why he went.

So your father had to pay $60 a year for a MENSA membership just to go to parties with (paid) strippers? Do you realize just how lame that sounds? Back in the 70s most of us who were adults then were going to parties with real sex with other participants, not strippers. We had a ski resort that the college rented over the christmas/new years holiday, that would put a student in charge (me at the time). We called it "The Last Resort." Student couples who wanted to get it on could stay there overnight - all it took was your student ID.

Hell of a lot more fun than a MENSA party with strippers. Your old man proves the point that MENSA is for lamers.

Donald Trump is rich and President of the United States, and you're not. Go figure.

Not from his own efforts. He got his old man's money and influence to start. And even then, he went bankrupt 6 (not 4) times. Same as many spoiled man-chilldren who never had to start from the bottom. And now he's bringing that talent to bankrupt businesses to government. His behavior has all the markings of a sociopathic con man mixed with the maturity of a 13-year-old.

Re:Smart enough to REALLY f*ck things up???

[mbkennel]

> Currently intelligence is not well defined

To the contrary, in the technical domain, the 'g' factor in the psychometric literature (you can call it 'general intelligence' or IQ) is well defined and has a quantitative meaning. In particular, it represents the empirically observed phenomenon that across a variety of cognitive/neurological tasks of different natures, there is a statistical tendency among humans for people who are pretty good at many of them have a much higher probability than background distribution of performing well on others. The tasks that don't correlate well don't get put into 'intelligence' tests, say like running and catching a fast moving ball.

The degree is how much this incorporates the aspects of the human word 'intelligence' in social situations.

Using the psychometric definition, it's clear that the machine intelligences can't even be measured because the background assumptions about correlations is false: 99.9999999% performance outlier on Go, brain-death on everything else.

Because machines today have no general intelligence---and it may not matter.

Re:Smart enough to REALLY f*ck things up???

[rtb61]

IQ is not a measure of intelligence it is a measure of learning ability, which is why it is tied to age ie the amount of time one has had to learn all sorts of stuff prior to the test (stuff in this case being the stuff in the test). It is not a measure of things like psychopathy or narcissism or belief maintenance (think belief in reproduction as an example, which is why religious nuts are such prolific breeders, even forced breeding through violence). There are a whole range of genetic traits which establish preferential thought patterns (think thought cycles repeated and tens of thousands of times a second https://en.wikipedia.org/wiki/..., which are actually a cyclic pattern on top of the molecular chain reaction which is much faster but still cyclic, think localised versus extensive) ie the direction in which thoughts will tend to flow due to cross correlation thought pattern tendencies as affected by brain chemicals (flow this way feel better, flow that way feel worse, feel good or bad is a really bad measure, feel better or worse is far more accurate).

AI is more a collective term for the application of computer learning algorithms, dependent upon task or element of task to be learned and cross correlated with out learning elements. Interesting thing in there, is what needs to be ignored and forgotten (getting an AI to ignore irrelevant data and delete unnecessary data, else AI in affect will go crazy ie lock up).

Re:Smart enough to REALLY f*ck things up???

[BarbaraHudson]

What you have described is not AI. When an AI can come up with a definition of what AI is, call me back. Until then, it simply doesn't exist, any more than filters that remove sludge from water are producing "intelligent water." Sorting and filtering are purely mechanical processes, same as they were back in the analog tube age.

Re:Smart enough to REALLY f*ck things up???

[rtb61]

Hmm, Intelligent water, the human brain, like duhh ;).

Re:Smart enough to REALLY f*ck things up???

[ChrisMaple]

Back in the 70s most of us who were adults then were going to parties with real sex with other participants

If you think that is intelligent behavior, it explains a lot of your posts.

he went bankrupt 6 (not 4) times.

Wrong. Corporations he ran went bankrupt. I don't know how many corporations he ran, or how many he had running at one time, so I can't estimate his success rate.

Your opinion of Trump seems to be based on his pre-Presidential public persona and the wailing of his enemies. Whether what he's doing is wrong or right, one thing is sure: he's got you fooled.

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

I keep hearing this repeated that IQ is not a good measure of intelligence. But almost every single test/study shows a very high correlation between high IQ and being good at activities that requires you to be intelligent.

It's like saying that math tests are not good at measuring someone's ability to do mathematics because there are some outliers that gets stressed when taking the test and underperform and some others that have memorized all the answers and over perform. Sure IQ test are not perfect indicators but if you claim that they don't show aptitude to solve logical problems, ability to remember things are speed at coming up with words starting with the letter R then you better show some proof. Because the academic world is full of examples where high IQ on a test correlates to high scores on other mental tests.

Re:Smart enough to REALLY f*ck things up???

[BarbaraHudson]

Hmm, Intelligent water, the human brain, like duhh ;).

Not really - the intelligence is in the impurities - they grey matter - not the H2O.

Re:Smart enough to REALLY f*ck things up???

[Anonymous Coward]

I don't think my mother would have appreciated him going to sex parties. She either didn't mind the strippers at MENSA meetings, or she didn't know about it.

When I buried him last year, I had to throw out his porn collection. Old horn-dog... Some of the hardcore stuff was autographed, so he probably had more fun than he let on. Stories I'll never know.

Re:Smart enough to REALLY f*ck things up???

[BarbaraHudson]

Given a choice between (1) paying $60 a year, then paying more for admission to a place to watch strippers while paying even more for each drink, or (2) showing your student IDs, getting a free room, bringing a reasonable amount of beer or other alcohol, and actually having sex with your partner, most people would choose (2) as the intelligent choice. More bang for less bucks, as contrasted to the MENSA more bucks for zero bang.

Re:Smart enough to REALLY f*ck things up???

[HiThere]

If you define "activities that requires you to be intelligent" as activities that require the things measured by an IQ test, then you have a valid, if nearly tautological, point. But there are a great many activities that accomplish useful ends which do no correlate well with IQ, and another cluster that only have partial overlap. E.g., changing a tire on a freeway.

So perhaps the problem is that we are using a different definition of which tasks require intelligence.

As for intelligence, I would argue that it requires intelligence to learn to play chess, but that it's not the same intelligence as the one that it takes to learn to deliver a speech. (Even that's too crude a partitioning, but if you finely partition things you start contemplating how intelligent a thermostat is ... not very, but a tiny bit, perhaps the smallest possible amount of intelligence. And this only makes sense in a very restricted set of circumstances.) And notice that neither chess playing nor delivering a speech correlates well with IQ.

Assumptions

[Anonymous Coward]

Nobody would ever misuse technology for financial gain, would they?

That's a lot of supersmart robots!

[Kargan]

We'd better get going if we are going to have more than 8 billion robots in 30 years! Right now, we have zero!

Re:That's a lot of supersmart robots!

[BarbaraHudson]

Depends on your definition of robot. Lower the bar enough, and my microwave, which cooks some of my meals, could be counted as a robot.

Same as if you scrape the bottom of the gene pool, you can easily conceive of "computers exceeding human intelligence" today, never mind 3 decades from now. We can start with the the guy quoted in the summary, Masayoshi Son, who doesn't understand the difference between computing power and intelligence.

Re:That's a lot of supersmart robots!

[m00sh]

We'd better get going if we are going to have more than 8 billion robots in 30 years! Right now, we have zero!

Robots do not have to be physical machines. They can be processes running on a device.

So, your current computer could run a robot process that does customer support, accounting or medical image diagnosis. There are a lot of jobs where the input and output are all inside a computer.

There are more than 8 billion devices. It could take a day or two to get all those devices the latest robot update.

Re:That's a lot of supersmart robots!

[Anonymous Coward]

Robots do not have to be physical machines. They can be processes running on a device.

Stop moving the goalposts.

Re:That's a lot of supersmart robots!

[drinkypoo]

Lower the bar enough, and my microwave, which cooks some of my meals, could be counted as a robot.

If it's got a temp probe or a pop sensor, then sure, it's a robot. For some reason I never have had one which has either, so I've never owned a robotic microwave oven.

Re:That's a lot of supersmart robots!

[Anonymous Coward]

Lower the bar enough, and my microwave, which cooks some of my meals, could be counted as a robot.

If it's got a temp probe or a pop sensor, then sure, it's a robot. For some reason I never have had one which has either, so I've never owned a robotic microwave oven.

If it has a timer, it's a robot.

Re:That's a lot of supersmart robots!

[tietokone-olmi]

We could just wait for people to become stupider than the average IoT doodad. Surely there'll be more than 8B of those by that time.

Re: That's a lot of supersmart robots!

[Anonymous Coward]

Robots are reprogrammable, flexible machines with at least 4 degrees of freedom that can function within a working envelope according to a program entered into a computer. A robot is not a Neural net, nor an AI, and those two are not equivalent. By the accepted definitions, especially industrial, a robot is a machine, not a program, and is separate from a computer which runs it.

Thus, fuck you, the guy being interviewed, and the horse you rode in on.

Re:That's a lot of supersmart robots!

[Maxo-Texas]

Read on up on quckly automobiles caught on and consider that many robots being sold today are fairly cheap.

Besides, an individual robot doesn't need to be better than every human at every task.

The 3,000 noodle chef's laid off in china by noodle robots are still out of work. And customers prefer the noodles created by the noodle robots (which.. keep in mind... cost less than the salary of a noodle chef in CHINA).

Re:That's a lot of supersmart robots!

[HiThere]

No. If it has a timer it has a bit more intelligence than a thermostat connected to a heater. But a robot needs to have the ability to manipulate things. So a toaster is a sort of minimal robot, but not a microwave, unless it opens it's door or pops up a switch (or rotates a knob) or some such.

Now what I'm trying to decide is whether that thermostat connected to a heater counts as a robot. It has internal moving parts, like the fan to blow the air. so it might be a sort of minimal robot.

At this point I feel like I'm trying to decide whether a virus is alive or not. I think by now the consensus is that it is, where it used to be that it wasn't, and what changed was not the virus, or even our knowledge of the virus (though that did change), but rather the definition.

Re:That's a lot of supersmart robots!

[HiThere]

I'd disagree about what you're calling a robot, though I'd agree that you're describing a 'bot. But we seem to be arguing about the definitions of words rather than about the thing being described. But this is significant if we each interpret the guy's predictions as being about our own meaning of the words. So with two reasonably common definitions we get either an unreasonable or a reasonable prediction about quantity of "robots", depending on which meaning we think he was using.

Re:That's a lot of supersmart robots!

[Anonymous Coward]

I have a robotic oven, microwave, dishwasher and washing machine. I also have a robotic AC and used to have a robotic heater. They all have built in AI. My car has AI fuel injection. Even my toaster has AI because it beeps 3 seconds before ejecting the toast. My water filter has an AI that tells me when I need to change the filters. Come to think of it, so does my refrigerator. I have a portable space heater whose AI keeps the temperature at exactly 20 degrees. I'm planning on installing a robotic watering system for my garden next year. I am already surrounded by dozens of AI robotic systems.

Re:That's a lot of supersmart robots!

[Anonymous Coward]

needs to have the ability to manipulate things

Like the flow of electrons?

alive or not

We had this discussion in my 5th grade gifted class

Re:That's a lot of supersmart robots!

[thinkwaitfast]

https://en.wikipedia.org/wiki/...

Re:That's a lot of supersmart robots!

[Anonymous Coward]

when I was a kid the school took us to the car factory, there the workers were making the cars, years after I returned for a visit, some areas didn't have any worker at all, no even for cleaning because there were self cleaning areas, no even lights on because the robots can operate with the lights off
now we have robots everywhere in industry, many in retail and slowly some even at home from the humble stupid washing machine to the robot vacuum cleaner, computing, information processing and sensors are cheap enough

Re:That's a lot of supersmart robots!

[HiThere]

To base the definition of "robot" on things that aren't perceptible to humans is probably a mistake. So I don't think that manipulating the flow of electrons would count. Logically I can't fault the argument, but practically English is based on human sensations. Purpose should also be a factor, though, so when talking about a computer the fact that it can eject a CD shouldn't make it a robot, but when talking about a CD Reader, there would be a reasonable argument. But we won't, because there's already a more specific name for the device. (This isn't a good argument, as a Roomba(tm) is clearly a robot, event though we could just call it a roomba, or even an automated vacuum cleaner.)

if we misuse it ...

[Anonymous Coward]

We will, rest assured .. mala tempora currunt sed peiora parantur

Re: if we misuse it ...

[Anonymous Coward]

Dont worry. The military think like u do and plan to misuse it in order to prevent u from ever using it.

Robot Buddy

[Anonymous Coward]

A robot buddy for every hikikomori! I want one now.

Enough

[Anonymous Coward]

Why the desire to keep reporting this crap?

"Up next: Elon Musk tells us about a dump he took."

Re:Enough

[Anonymous Coward]

"Up next: Elon Musk tells us about a dump he took."

Holy Shit.

Re: Enough

[Anonymous Coward]

Praise him!

Re:Enough

[powerlord]

"Up next: Elon Musk tells us about a dump he took."

Holy Shit.

No, you're thinking of "The Pope Just Took a Dump and You Should See What the College of Cardinals is Thinking"!

Elon Musk's Article was "Amazing Shit"

Re:Enough

[Anonymous Coward]

"Up next: Elon Musk tells us about a dump he took."

Holy Shit.

Apparently it took off and then landed itself on the platform all by itself! The best nasa could do is light one off and have it land in the water!

Re:Enough

[Anonymous Coward]

And he thinks it doesn't stink.

I'm going to assume...

[Type44Q]

I'm going to assume that the "E" in this clown's title does not stand for "engineer." :)

not difficult.

[Anonymous Coward]

...will exceed human intelligence...

A remarkably low bar.

Re:not difficult.

[Anonymous Coward]

As if there is something more intelligent that isn't human.

It really speaks to your lack of intelligence if your "low bar" is equal to the highest level of intelligence that we know of. Even using someone as retarded as you as for the baseline for "low bar" it is still higher than all other animals. That isn't a compliment. You're still retarded.

"If we misuse it, it will be a risk.

[Anonymous Coward]

If we use it right, it can be our partner."
So, pretty much like everything?

'cept maybe cigarettes...

Then the Cylons will revolt

[jfdavis668]

And we will have to escape on the nearest Battlestar.

Re:Then the Cylons will revolt

[Anonymous Coward]

We're not going to have a Battlestar. We're going to have a wall.

Re:Then the Cylons will revolt

[jfdavis668]

Great! That will keep the Cylons out.

Just do not connect them to GOOGLE

[p51d007]

We don't want them to be self aware ;)

Re:Just do not connect them to GOOGLE

[glenebob]

I think it's probly more important that they don't have opposable thumbs.

Oblig. Global Warming

[pipingguy]

Who's going to be around in 30 years to check if this guy was correct? We'll all be dead from ClimateChange(TM) by then!

Tip: Always make extraordinary predictions far enough into the future that no one will be around to verify (or they won't remember). Also, make as many predictions as possible and downplay the ones that turn out to be 'inconvenient'.

Right in the jerbs!

[Anonymous Coward]

Indo chimp robots are taking our jerbs!

Moore's Law continues?

[Anonymous Coward]

Power consumption per unit of processing power keeps going down too? I don't think these people know what is at the core of the thing they're extrapolating. Shoes will definitely not outsmart us.

Re:Moore's Law continues?

[glenebob]

If shoes have computers in them, at some point it will be cheaper to put computers in them that are smarter than us. Even a fairly dumb shoe understands this.

Re:Moore's Law continues?

[Anonymous Coward]

I thought we are now near the limits of UV lithography. How are computers going to become much smarter without new tech?

Re:Moore's Law continues?

[Anonymous Coward]

If shoes have computers in them, at some point it will be cheaper to put computers in them that are smarter than us. Even a fairly dumb shoe understands this.

Just strap two out of work Indian IT "consultants" to your feet and yell "Mush!" at them when you want to engage "Walk
" mode.

Pros: Cheaper and smarter than new collectible status symbol grade Air Jordans
Cons: Compared with basic used sneakers, they smell worse in summer.

Re:Moore's Law continues?

[ChrisMaple]

Single atom placement has been an established technology for many years, it's just not practical because it's very slow.
Obviously, once we get to wavelengths so short that we can't focus them and selectively block/pass them, we're going to need different technology, so at some point the word "lithography" may have to be dropped or its meaning changed.

Do take a look at wikipedia articles on extreme UV lithography, X-ray lithography, electron beam lithography and ion beam lithography.

bang the gong

[CresCoJeff]

do they... do they let just anybody say whatever they want at these things? Dude's making a nutter claim with zero sources cited, and isn't even in the AI field himself... and Fortune picks it up like it's totally reasonable. Also, if the whole "we will be less than our shoes" thing wasn't a translation error, Son's got a troubling conception of what makes a given life form 'worthy' (i.e. raw IQ is directly proportional to worth, and nothing else matters)

The banality of ubiquitious genius will doom us

[quietwalker]

We have people barely able to tie their shoes who get bored at their low/no-skill minimum wage job now, and they're going to be the first to be replaced. What's going to happen when we turn over their jobs to super smart AI-powered machines? Are fast food order kiosks gonna be the start of the robot uprising? ... and what a boring way to begin a sci fi novel: "Day 1 of the robot uprising: exactly 13.74% of the McDonalds orders for large sodas were substituted with medium sodas, a precise amount calculated to cause the maximum dissatisfaction without rising to a level where we would be alerted. We didn't know it, but it was already too late. They had already calculated every possible move. On Day 2, there was nothing to stop them from adding pickles to orders that expressly asked for no pickles. It was the end times."

Re:The banality of ubiquitious genius will doom us

[TrumpShaker]

LOL.. nice thought, but what if the percentage of correct orders is still higher than before with what you envision? Day 3, replaced workers return with a vengeance (and baseball bats) and yep...the end times, first day of the war...

Re:The banality of ubiquitious genius will doom us

[m00sh]

We have people barely able to tie their shoes who get bored at their low/no-skill minimum wage job now, and they're going to be the first to be replaced. What's going to happen when we turn over their jobs to super smart AI-powered machines? Are fast food order kiosks gonna be the start of the robot uprising? ... and what a boring way to begin a sci fi novel: "Day 1 of the robot uprising: exactly 13.74% of the McDonalds orders for large sodas were substituted with medium sodas, a precise amount calculated to cause the maximum dissatisfaction without rising to a level where we would be alerted. We didn't know it, but it was already too late. They had already calculated every possible move. On Day 2, there was nothing to stop them from adding pickles to orders that expressly asked for no pickles. It was the end times."

No. The no-skill isn't going to be the first to be replaced. They are so low wage and involves manual labor that it would cost a huge investment to replace them.

The first ones to be replaced will be mid to high skills jobs where the output is just a computer file and the input is some human input. For example, some types of lawyers, parts of customer support and some aspects of medical service. And, dare I say coding?

The reason that it will be first to be replaced is because once the software is ready, there is almost zero infrastructure cost. Rent a VM in the cloud and any tablet/computer can be a job replacing robot. You don't need sophisticated computer vision mechanical robots needed to flip burgers.

Re:The banality of ubiquitious genius will doom us

[AmiMoJo]

In Larry Niven's Puppeteer society nothing that can be done by citizen is automated. There are too many citizens, they live too long and they get bored, so would rather do menial tasks than nothing.

A more likely scenario for humans is that there will be a difficult period of transition, where people are used to working and being paid and have to adapt to living on some form of welfare (e.g. universal income) and very early retirement. Those who do carry on working will probably be resentful and angry, even though they don't really want to stop working or get less money.

Eventually people will only work if they choose too, and many will choose it. Robots will do everything else. Probably super smart ones will make lesser robots and control them, because wouldn't you make a dumber clone just to do all the drudge work in your life while you attend to the interesting bits.

Re:The banality of ubiquitious genius will doom us

[Anonymous Coward]

The first ones to be replaced will be mid to high skills jobs where the output is just a computer file and the input is some human input. For example, some types of lawyers, parts of customer support and some aspects of medical service. And, dare I say coding?

Robots will never replace coders. We had a poll on the issue, and the results were final.

Re:The banality of ubiquitious genius will doom us

[TrumpShaker]

Ironic, related links at the bottom, first one says: "Former McDonald's USA CEO: $35K Robots Cheaper Than Hiring at $15 Per Hour"

Re:The banality of ubiquitious genius will doom us

[Anonymous Coward]

low/no-skill minimum wage job... they're going to be the first to be replaced

What makes you say that? Seems like the easiest jobs to automate would be managerial roles -- the ones where most of the "work" amounts to decision-making. Those are the jobs that software alone could replace. Minimum-wage jobs, on the other hand, don't involve much decision-making. So unless their tasks are easily standardized and automated with robotics, they're not going to be replaced any time soon.

Re:The banality of ubiquitious genius will doom us

[Anonymous Coward]

Right. This will be a very unpopular opinion here, but replacing an industrial line worker that does slightly more than a mechanical process is much more difficult to replace than white collar knowledge workers.

I was employed by an automotive supplier that manufactured goods for BMW, Lotus, Ford, Toyota, and piecemeal for others. A single line worker that oriented, fixated, ran processes on a part and then did initial QA was replaced by several robots, cameras, automated equipment stations, and several computers. Each replacement "cube" (as that is what shape they ended up being) cost just short of a million dollars. A large part of my job was fixing the buggers no matter how long it tool when they failed to do their job. If only we treated humans so well.

Meanwhile the entire accounting department of half a dozen was replaced by just one module in the EMS system that cost 12k. The human element was outsourced to a CPA that came in for one week out of the year. The HR and sales departments were cut from 20 down to 3 with an opensource CRM package. To run all the new software not one new IT person was hired.

Replacing the $10/hr workers cost a million each. Replacing 3 million dollars in salaried employees cost 20k. The only saving grace that white collar people have had was that they got to decide who was cut first.

Re:The banality of ubiquitious genius will doom us

[Anonymous Coward]

Stop causing static!

Re:The banality of ubiquitious genius will doom us

[Shotgun]

A cloth mill down the road used to employ thousands. Now there is a guy to occasionally deliver new thread bobbins. The machines load the thread themselves.

Most of the easy to replace manufacturing jobs have already been replaced. The fast food workers will be replaced when they demand a $15/hr minimum.

I write software test automation. I've written scripts over the last three months that replace what 6 people did in a year. On the bright side, I have seen my salary increase 25% over the last two years.

Re:The banality of ubiquitious genius will doom us

[SuricouRaven]

I'm expecting to see automation lead to huge inequality for a time, with a large underclass who have no employment and little hope of employment, and a working class who are heavily taxed to pay for them and resent that their hard work is being stolen by a bunch of useless leeches.

Eventually it goes one of two ways: Either the leeches revolt and bring about some form of revolution which may or may not result in a working economic system, or the government has to evolve into a police state in order to suppress the frequent riots that arise from having a very poor, very angry population. Especially one with little else to occupy their time. The masses may be placated with the bare minimum of resources needed to keep them from mass-starvation and a steady supply of entertainment, but the upper classes will resent even that much.

Re:The banality of ubiquitious genius will doom us

[pr0fessor]

cashiers... there are self check outs all over with one 1 cashier monitoring 6-10 check out lanes, you don't need to go into the gas station any more you just swipe your card at the pump, and ATMs are all over the place as well. My town has gone from two video rental stores to a dozen redboxes.

Re:The banality of ubiquitious genius will doom us

[thinkwaitfast]

The lawyers won't allow this to happen

Re:The banality of ubiquitious genius will doom us

[Anonymous Coward]

Have you ever tried to ring up a full shopping cart of groceries on one of those self-checkouts? While you may of got through this, how many times were you held up because numerous other people just couldn't figure it out?

The grocery store closest to my home just took the self checkout OUT of the store and is putting in two more lanes. I really hope they increase the payroll for the frontend along with it.

I'm imaging the store was losing more money from intention and unintention theft then it would of cost just to run additional checkers. A good checker can really speed things up, especially for elders that cannot physically move faster if they had to.

Re:The banality of ubiquitious genius will doom us

[ChrisMaple]

Settled science, just like global cooling --- er, global warming --- I mean, climate change.

Re:The banality of ubiquitious genius will doom us

[ChrisMaple]

Never been a manager, have you?

Re:The banality of ubiquitious genius will doom us

[ChrisMaple]

Self-checkout seems to have some limiting factors and in some businesses has already reached saturation.
A supermarket near me closed down its 4 self-checkout machines and returned to human cashiers.

Re:The banality of ubiquitious genius will doom us

[pr0fessor]

A walmart near me just replaced 4 express check outs and 2 regular check outs with 8 more self check outs for a total of 16. I have no idea what factors are pushing the self check out in some areas but not others. We have a couple fast food places with order kiosks that I doubt will last but the self check out in walmart and the local grocery are always packed along with the unattended gas station {not legal in all states}.

Re:The banality of ubiquitious genius will doom us

[Anonymous Coward]

You're missing the obvious third alternative: automated weapons simply killing off the underclasses.

Re:The banality of ubiquitious genius will doom us

[quietwalker]

Wrote about this several times before. To steal from one of my previous posts...

I'll summarize it for you all though. In order to avoid a situation where the majority are unsustainably poor and ready to revolt, we'll need to meet the following criteria:
        - Every country in the world needs to be at about the same technological level at about the same time
        - Every government in the world (and all the people within them) embrace strong socialist beliefs that make current socialist states look like anarchists
        - We need to abolish the concept that work is directly related to value, and in turn, diminish the concept that scarcity and demand have real impact on value.
        - We have to accept that there is going to be a sizable number of people in the world who add no value to society or the world, and simply exist as consumers

The average person would have a trade skill that they use when they feel like it, perhaps no more than 1-3 hours a week, live in a house or home they like, and their things (clothing, devices, transportation, food etc) would be freely given to them with only limits placed on quantity by need - for example, no one needs more than 1 car, but you might - from time to time- need a truck or a motorcycle. There'd be no such thing as money, private ownership of property (items & land) is almost completely gone, and naturally limited resources would be metered out by some merit plus popularity based system, so not everyone would have a starship, for example. ... but the reality is that we're probably going to have to go through at least one, if not more cycles of horrific violence or strife, to the point that it forces us to radically redefine our thoughts and behaviors. We're just too caught up in concepts of justice-as-defined-by-the-beholder, us-vs-them, and so on to do it right now.

30 years... no

[fyngyrz]

It almost certainly won't be 30 years.

It'll be 1-3 years after the first one appears.

Re:30 years... no

[avandesande]

It will be 5 years after the first commercial nuclear fusion plant is started.

Re:30 years... no

[Anonymous Coward]

is that before or after i get my flying car.

Re:30 years... no

[Tim12s]

I guess, approximately 12 years after unlimited energy and unlimited manufacturing capability across all segments of the economy, we will have the first generation of kids with zero ability to compete and therefore the possibility economic collapse. It is also likely that the economic implications will not be fully comprehended and incorporated into the world economic system which means that specific corporates are competing. Multiple system shocks will likely put the worlds economic system at a breaking point, after which, based on mankind's ability to work together, it will be a choice between (a) nirvana / star trek / culture, or (b) dystopian future / enslavement by the few, (c) the apocalypse / terminator .

Personally I prefer (a). Our ability to work together has not been proven. Consider Kim/Bob/ with unlimited energy and unlimited manufacturing.

Re:30 years... no

[Anonymous Coward]

Just you wait. Within our lifetimes, we're going to see an Artificial Intelligence crash its flying car into a commercial fusion plant and blow us all to hell. And if we survive, you'll have egg on your face for all this mockery!

Re:30 years... no

[AK Marc]

Why would there be more than one? The first would be the only one, and would take over the world as the only AI. Unless two AIs were created separately and isolated concurrently, but they'd likely battle as soon as they were aware of each other.

Unless the AI makes slave AIs because it doesn't "want" to do something and can't trust dumb humans to do it.

Re:30 years... no

[pkwatz]

Mandatory XKCD: https://xkcd.com/678/

Re: 30 years... no

[Anonymous Coward]

Actually, we will have a human *almost* do that and an AGI will save them. And it is going to happen a heck of a lot faster than anyone realizes. We will *all* be out of jobs in 5 years. Every last one of us, from physicists to burger flippers and everyone in between. And we are so stupid, we just let Russia "elect" the one U.S. President that will take 99.99% of us in the wrong direction... towards abject poverty with absolutely no lifelines. It is mass extermination. There will be *no* jobs for programmers.

Unfortunate data set choice...

[Anonymous Coward]

Unfortunately, their choice in data sets was looking at the rate of change in world leader intelligence over the past couple of years.

By that measurement, human intelligence will be overtaken by jellyfish intelligence by 2030.

Well, actually, in light of the use of world political and economic power, what robotic intelligence we do get may be shaped largely by the purse strings of those garnering all the power, largely fossil fuel brokers and those that serve them. Since it's a lot easier just to surround people with a blanket of inescapable lies than actually develop convincing intelligence,

I'm expecting we'll all get our own personal Kelly-Anne Conway using a simple ELIZA loop fed with talking points, rather than true AI for a LONG while, since apparently, that's all that's required to get what folks really want.

So 9 billion robots ...

[CaptainDork]

... by then.

I doubt that.

Also, recall CaptainDork's famous corollary: "For every smart mother fucker out there with a computer, there's smarter mother fucker out there with a computer."

No entity is smarter than its creator.

God told me that.

I'd show you, but he used Snapchat with 20s TTL.

Re:So 9 billion robots ...

[CrimsonAvenger]

Not sure where the "outnumber" in the title of TFA came from. The link never bothered to mention the number of robots, though it might have been sorta implied by the comment about our sneakers having more intelligence than us by then...

If I were guessing, I'd guess someone mistranslated a word, and out "editors" didn't catch the mistranslation....

Re:So 9 billion robots ...

[Anonymous Coward]

No entity is smarter than its creator.

That doesn't say much about your parents.

Re: So 9 billion robots ...

[Anonymous Coward]

That's got me thinking. Is it "motherfucker" or "mother fucker"? I've always used the former.

Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[Rick Schumann]

I'm getting a bit sick and tired of these bullshit stories. Robots are not going to make humans obsolete. Everyone is not going to lose their jobs to robots. Robots are not going to take over. So-called 'AI' as most people THINK they know it is a creation of media hype. Everybody needs to calm the hell down and just go about your business, nothing to see here, ignore the trolls and the bullshit 'news' stories.

Re:Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[Oswald McWeany]

If we continue improving our abilities, and don't wipe ourselves out, we will absolutely produce a superior-to-human AI one day. I couldn't give you a time frame, and neither can anyone else realistically, but one day it will happen. My lifetime? My kids lifetime? No-one really knows, it's inevitable one day assuming we don't wipe ourselves out some other way first.

Re:Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[Rick Schumann]

We don't even understand how our own brains work yet, let alone how we're conscious. Until we know that we aren't going to have 'real' artificial intelligence. I talk to people who research these things, that's what they tell me, and they also say that all the media hype is just that: hype. Meaningless. They're mis-using the term 'artificial intelligence'. The average person (media people included) actually believe the fantasies that you see portrayed on TV and in the movies; we're nowhere near any of that, and won't be for a long, long time to come. So as I said above: Everybody needs to CALM DOWN. Like I keep saying: Don't talk to me about 'artificial intelligence' until you can bring me something I can have a full-on, spontaneous, completely random, unscripted, and completely coherent conversation. When that day comes I'll call that thing 'AI', and not before. I'm talking Turing Test, and beyond. Anything short of that is just a poor imitation and I'm not impressed.

Re:Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[SuricouRaven]

It doesn't take everyone losing their jobs to robots to really screw up the economy. There's a feedback system to consider. If, say, 10% of the workforce lose their jobs to automation, they can no longer afford to buy nice things, which means less economic activity to generate employment for the rest of the population. Just look at what happens to the economy of any country during a recession, and the resulting implications for individuals caught within it. Automation can be seen as a recession that never ends, because even when productivity goes up employment barely moves.

Re:Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[Rick Schumann]

Sure, that all sounds perfectly reasonable and logical. But the alarmists with their sky-is-falling rhetoric are full of crap. Corporations might well choose profits over people, but governments will not. If they don't then there will be riots and maybe even civil war, since people are not going to just sit placidly by and starve to death. Governments will encourage retraining of people, and/or make it less attractive for corporations to fire humans. You can call it 'protectionism' all you want but when it comes right down to it, human beings cannot become 'obsolete', ever, and it's absurd to suggest it. Without humans there is no reason for any automation to exist.

Re:Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[Anonymous Coward]

We don't need to understand how our brains work to make an AI, we don't even need to understand what goes on inside the brain of an AI to make one
We make (allegedly) intelligent people every day effortlessly and we don't even understand how that happens
It wont be the firs time we invented something that we didn't understand why/how it worked until many years after

Re:Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[Oswald McWeany]

There's already chat bots that can fool people some of the time, progress, naturally would imply that improving. Those aren't really "AI", but certainly I expect chat bots that can pass a Turing Test within a decade (and that's not a conservative guess), we're already moderately close. Those aren't really AI though.

Real AI, I'm not saying it's coming any time soon, I really don't know. I'm just going to say it's inevitable if we don't wipe ourselves out first. We move closer every year and certainly we have attributes of AI even if we don't have AI yet.

I remember having conversations where they said there would never be an electric car on the street because it would be too expensive, and we it is scientifically impossible to have an electric motor perform like an engine. You'll never get enough range for the average consumer... etc,etc, etc.... we all know now that that isn't true, electric cars are growing quite rapidly. That was less than a decade ago. As recent as two years ago I remember people arguing that you will never have self driving cars, it was a fantasy technology no-one would want, Tesla is essentially doing that today too!

Then there is the whole "we'll never get much more data density than a CD in an optical drive because of laser wavelength".

So AI? Yes, it can be done, there is no reason we can't duplicate what is done with cells in our brain today with some form of computing (we already have more computing bandwidth than the brain, we just haven't figured out the abstract though). It probably won't be in 10 years... it might not be in my lifetime, or even my kids, but, unlike other technologies, it is obviously feasible, and there is an obvious desire for one. It will happen if we don't wipe ourselves out (or our technical ability out), it's when, not if.

Re:Fear! Uncertainty! Doubt! EVERYBODY PANIC!!!

[RespekMyAthorati]

I expect chat bots that can pass a Turing Test within a decade (and that's not a conservative guess),

No, it's a baseless guess.

we're already moderately close.

NO, we are not "moderately close".

Every year, there are chat-bot contests, and every year, the winner is
a total joke that answers difficult questions by changing the subject.
Where they find these idiot testers, I can't imagine.

We need an AI for filtering Hype

[tomxor]

This is ultra hype... AI today is both very powerful and very stupid, people who do not understand this create this extreme hype... it's not even correct to call it stupid because in the low level functionality of AI today "clever" doesn't exist (that's built into the upper layers that don't exist in artificial NN).

The latest AI is powerful because we can directly manipulate the design of a relatively tiny network that do relatively basic things, making them do what we want can be attributed to the cleverness of the computer scientists creating them to do specific tasks, not the resulting network it self... This is where the hype begins, there is a galactic sized space between this infinitesimal functionality and the network of networks of networks of networks of networks that make up something remotely conscious and "smart" in the inventive, creative and insightful sense that we think of humans being smart.

The best analogy I can give is to say that AI today are individual bytecode instructions, the building blocks... consciousness is the operating system that grew on top over a billion years of iterations.

Disclaimer i'm not a computer scientists studying AI, but I've read enough to separate the hype from the substance.

Re:We need an AI for filtering Hype

[religionofpeas]

The latest AI is powerful because we can directly manipulate the design of a relatively tiny network that do relatively basic things

The funny thing is that the latest AI isn't fundamentally different than the AI in the 80's that didn't work. We've made a few small improvements in the function, but most of the progress has been made by increasing the size of the networks.

This is where the hype begins, there is a galactic sized space between this infinitesimal functionality and the network of networks of networks of networks of networks that make up something remotely conscious and "smart" in the inventive, creative and insightful sense that we think of humans being smart.

It's possible that we only have to make a few more small improvements, and a further increase of the network size to achieve that. That's how our brains evolved from small rodents to us.

CEOs are smarter than anyone

[TiggertheMad]

Well, since a CEO thinks that this will be true, it must be. I love how CEOs like this guy and Elon (idiot) Musk are predicting the future of AI development. As opposed to say, leading AI researchers that are attending conferences and writing papers on the state of the art.

My response: STFU register biscuit, and work on growing your companies valuation rather than talking about shit that people way smarter than you cannot predict. This headline might as well be, "Random unqualified person speculates on the unknown future".

Re:CEOs are smarter than anyone

[ShanghaiBill]

As opposed to say, leading AI researchers that are attending conferences and writing papers on the state of the art.

The experts in any field tend to be focused on the problems and obstacles, and are often the unduly pessimistic about progress. In hindsight, they often turn out to be the worst predictors. It is hard to see the horizon when you are in the trenches.

Re:CEOs are smarter than anyone

[khallow]

I love how CEOs like this guy and Elon (idiot) Musk are predicting the future of AI development. As opposed to say, leading AI researchers that are attending conferences and writing papers on the state of the art.

What do AI researchers have to do with AI? Let's recall some things you've said before:

Ten years out? As a veteran programmer and AI enthusiast, I'd say it was more like a century. We cannot build a computer that can model a bug's brain activity, let alone something a million times more complicated like a human brain. And that doesn't even get us to the 'superhuman intelligence' category that people are afraid of.

In other words, AI researchers are in your opinion a century out from having a relevant opinion on AI. Then there's your discussion with a MS game developer:

I was having a conversation with a guy who was working on AI algorithims, and I asked what sort of schemes he used, fuzzy-logic, Genetic learning, or weighted neural nets? He told me that they didn't bother with academic AI techniques, because he could already write an AI that could beat the player every time without them.

I was completely at a loss for words, so I just thanked him and ran away.

Yet another indication that the "academic AI techniques" might not be up to snuff for dealing with real world AI creation and consequential issues.

Re:CEOs are smarter than anyone

[PCM2]

I love how CEOs like this guy and Elon (idiot) Musk are predicting the future of AI development. As opposed to say, leading AI researchers that are attending conferences and writing papers on the state of the art.

Hmm. You mean that same "idiot" who founded an artificial intelligence research organization to help fund the very things you hear about at those conferences and in those papers?

Re: CEOs are smarter than anyone

[Anonymous Coward]

So he hired a bunch of people to work on the thing he knows nothing about. Wow that's some great proof you got there. We will all be replaced by smart cat robots in no time for sure.

Re:CEOs are smarter than anyone

[Anonymous Coward]

As opposed to say, leading AI researchers that are attending conferences and writing papers on the state of the art.

The experts in any field tend to be focused on the problems and obstacles, and are often the unduly pessimistic about progress. In hindsight, they often turn out to be the worst predictors. It is hard to see the horizon when you are in the trenches.

Experts with "book learning" are the worst! They keep trying to dismiss my perpetual motion and free energy machines as cranks! All I need is for someone to make the parts the way I tell them to from perfectly elastic, massless, frictionless and unbreakable material!

Re:CEOs are smarter than anyone

[Anonymous Coward]

Nonsense. Link? Do you mean like when Einsten and a slew of physicists wrote Roosevelt a letter asking to fund the Manhattan Project? The all knew YEARS before anyone else did that nukes were possible. Or do you mean like when people were building rockets 30-40 years before Apollo 11 fully convinced that the would get to space or the moon in their lifetimes. Sorry, but people in the sciences know the real state of things. That hasn't changed. What has changed is the level of incentive to hype total BS for funding. That's never been greater.

Re: CEOs are smarter than anyone

[Anonymous Coward]

I tend to agree, except. Deep learning sucks, it just .... sucks. Until such a concept is even integrated into our lives functionally AI is and will be just a other pipe dream.

If you understand the very basics of computer science you realise concepts spanning back decades form the basis of the most advanced technologies today. The fundamental limitations then vs now are still essentially present, the key differences to are that of size and energy efficiency but having said that we still have plenty of a way to go.

Hence why we're headed for another tech bubble as the KoolAid is still running deep, lots of idiots out there, just like this guy.

Re: CEOs are smarter than anyone

[Samantha Wright]

Bingo. All of the new and exciting developments of the last decade have been in machine learning tasks. That says nothing about generalization. What we're developing are very efficient tools to accomplish new tasks, but those tools have precisely zero skill in the cleverness department. It's so common for people to forget this fundamental distinction that there's a term for what happens when they remember it. AI research has its own genre of tech bubbles caused by overoptimistic futurists.

Re:CEOs are smarter than anyone

[mbkennel]

> The experts in any field tend to be focused on the problems and obstacles, and are often the unduly pessimistic about progress. In hindsight, they often turn out to be the worst predictors. It is hard to see the horizon when you are in the trenches.

Do you have widespread examples about this?

Re: CEOs are smarter than anyone

[Anonymous Coward]

I'd have said experts in AI have tended to be overly optimistic, and I say this as someone who worked on neural networks.

Re:CEOs are smarter than anyone

[AK Marc]

As opposed to say, leading AI researchers that are attending conferences and writing papers on the state of the art.

AI researchers spend more time weakening the definition of AI than delivering any progress towards strong AI. Back 50 years ago, "AI" meant "Strong AI" and it's only recently when we realized AI is hard, so we'll use AI to mean "anything that seems smart" and deliver useless things and claim a victory.

Re:CEOs are smarter than anyone

[constComment]

Actually, research has shown that pessimistic people tend to have a more accurate grasp on reality. (Eg. Eighty percent of small businesses fail. Pessimists will avoid this like the plague. Only optimists will attempt something that is most likely not in their favor.) So, if researchers are indeed pessimistic (I do not know though that is a given) -- I will take their opinion over that of a CEO.

Re:CEOs are smarter than anyone

[GuB-42]

Experts are better predictors, that they appear bad is just a statistical effect.
It is just that experts tend to be more consistent in their predictions because they know their stuff. OTOH amateurs are more likely to go wild, and make more exciting and varied predictions. It means that while experts are more likely to be right, the one with the right answer is more likely to be an amateur, and if it isn't, it may not be exciting enough to make it to the news.
If you want an analogy, imagine a loaded dice, where number 3 is twice more likely to come up than the other numbers. Experts, knowing that the dice is loaded, will predict a 3, others will predict any number, with a preference toward exciting numbers like 1 or 6. Roll the dice, 2 out of 3 times, the experts are wrong and some amateurs will be right, even though experts picked the most likely choice. If we only talk about 1 and 6s, it will appear that the experts are always wrong while plenty of amateurs are right.

Isn't all of this just BS?

[Eloking]

I'm I missing something or all of this (news about AI taking over) is just BS?

Almost as bad as a "Terminator" type of rebellion : https://what-if.xkcd.com/5/

Yeah we see a lot of breakthrough in "AI" technologies (AI beat GO champion last year, AI got better to identify skin cancer this year), but as far as I understand AI, it's basically plugging the program to a (insanely huge) database about the subject and help him interpolate the input and it's own data. That's computer program getting better, not getting "intelligent".

Or is my definition of "AI" that off the mark? I mean, for me intelligence implies some sort of "conscience" that can make decision "outside the box". No matter how fancy the GO of dermatologist AI get, they will never do more than their field because they are not programmed to do so.

Re:Isn't all of this just BS?

[religionofpeas]

I mean, for me intelligence implies some sort of "conscience" that can make decision "outside the box".

People don't make decisions outside the box. Some people just have a big box.

Re:Isn't all of this just BS?

[phantomfive]

but as far as I understand AI, it's basically plugging the program to a (insanely huge) database about the subject and help him interpolate the input and it's own data

Alphago is also a huge tree-searching algorithm, with tremendous processing power. The real question is, how do you know your brain does something more advanced than that?

Re:Isn't all of this just BS?

[Kjella]

as far as I understand AI, it's basically plugging the program to a (insanely huge) database about the subject and help him interpolate the input and it's own data. That's computer program getting better, not getting "intelligent". Or is my definition of "AI" that off the mark?

Well it depends on how much you consider "MacGyver" style problem solving to be intelligent. As in I have a task to complete, I have a bunch of random items that can be combined/used in some way to produce a non-obvious result. Computers are great a combinatorics even to the point where they might do something that's original and never been done by a human. A lot of what humans consider creative is putting together known things in unexpected ways, or at least that this particular person has never done before. You might say that the computer is always in the box but we're trying to expanding it while at the same time guiding it so it doesn't get lost in an endless number of possibilities.

Maybe it's easier to explain with a practical example, before you gave the computer a toolbox and taught the computer that the the hammer could hammer, the saw could cut and the screwdriver screw and that was the box. Then we gave it free roam as a few hunks of wood and metal and it got totally lost. Now we give it examples of people hammering and cutting and screwing which guide it, but doesn't bind it. And we find that sometimes it does things in novel ways because nobody told it that it couldn't. The goal is to make "the box" the laws of nature, physics, chemistry, gravity, optics and so on. That we stop defining for the computer what something is and what it can do.

Re:Isn't all of this just BS?

[thinkwaitfast]

AI's were beating people in tic-tac-toe in the 1960w.

Re:Isn't all of this just BS?

[thinkwaitfast]

Can alphago play chess or would if have to be reprogrammed? Would it know what it know what to do if you gave it a tictactoe board? Or risk? Empire builder? D&D?

Re:Isn't all of this just BS?

[phantomfive]

You would have to teach it the rules, but the same algorithm would largely work.

Re:Isn't all of this just BS?

[Eloking]

AI's were beating people in tic-tac-toe in the 1960w.

Well yeah, but anyone with a minimum experience can play equally and not lose. There's only that many combinations of tic tac toe.

If you want an early field where computer destroyed the human, you best bet will be the calculator.

Re: Isn't all of this just BS?

[Anonymous Coward]

No. The real question is, can it decide it doesn't want to play?

knowledge vs intelligence

[swell]

These 'artificial intelligence' devices are trained by giving them examples so that they can learn to recognize certain items/situations, etc. Show a neural net 100,000 pictures of various dogs, and eventually it will be good at recognizing pictures of dogs. Thus they acquire knowledge, but have they become intelligent?

Intelligence involves responding appropriately to a novel situation. Extrapolating from often vastly different experiences in some cases, or simply total innovation. This is what intelligence tests attempt to discover. Ideally they require no previous experience with similar questions/puzzles. They simply require mental agility of a kind that computers have not been good at.

Re:knowledge vs intelligence

[swilver]

Show it an upside down dog...

Re:knowledge vs intelligence

[swell]

... and even when the computer learns to recognize a dog by its appearance, it will not grasp the sound, the smell, the feel of a dog, or the difference between a snarl and a wagging tail. The computer will never know or understand the warmth and loyalty of a dog, or suffer pain at the loss of a dog.

Show me your super smart AI now!

No

[Artem S. Tashkinov]

Computers running artificial intelligence programs will exceed human intelligence within three decades

No. More like in 300 years considering the "progress" in AI. We are good at creating highly specialized algorithms for completing certain tasks. We have nothing generic which can function and solve never seen before tasks completely on its own.

What's more we still have no idea what intelligence and consciousness are. AI is certainly a buzzword today considering the number of recent films dedicated to it. But something tells me we'll soon hit another major winter in AI development unless someone comes up with better ideas about intelligence or its implementation in silicone because we're quickly running of ideas and transistors - Moor's "law" (there has never been any law, except in the imagination of some silly journalists) is barely alive and it'll soon hit the wall of physics and diminishing returns.

Re:No

[religionofpeas]

No. More like in 300 years considering the "progress" in AI.

The progress on AI in the last 10 years was more than the 290 years before that. And Moore's law is far from dead. Just look at what our brain can do, and they're made from milk and sandwiches.

Re:No

[Baron_Yam]

Scientists have already built an hippocampal prosthesis. https://en.wikipedia.org/wiki/Hippocampal_prosthesis

You might also look into memristors, which show some promise as artificial neurons far more effective than simulating them with transistors.

Neurology is also starting to crawl out of the dark ages, and a renaissance there will immediately spill over to AI research.

It may turn out that we don't have to figure out how it works if we can 'simply' build a replica of a biological model in silicon... though I'd suggest that might be less than ideal as we'd probably want to understand how to program a mind before creating one that would be damn near functionally immortal and could operate faster even if not smarter than ours.

My point is that an AI breakthrough is somewhere between a decade from now and 'forever', but I wouldn't set any particular minimum with a high level of confidence.

Re:No

[phantomfive]

The progress on AI in the last 10 years was more than the 290 years before that.

I don't think that's true. The algorithms we are using were mostly invented before 2007. Deep learning merely means taking those algorithms and applying incredible computing power to them.

And Moore's law is far from dead.

That seems to be true. Although CPUs have been largely stagnant, GPUs have been jumping dramatically in performance still. And that's where the processing power comes from.

Re:No

[phantomfive]

It may turn out that we don't have to figure out how it works if we can 'simply' build a replica of a biological model in silicon

The chances of building a mind without understanding how minds work seems lower than building a functioning rocket without understanding physics.

Re:No

[religionofpeas]

The algorithms we are using were mostly invented before 2007. Deep learning merely means taking those algorithms and applying incredible computing power to them.

I didn't mean to imply that the algorithms were novel, just that the results have exploded. And now that we know we're on the right path, progress will continue by exploiting specialized hardware for this particular function, instead of using general purpose computers.

Re:No

[religionofpeas]

Our minds have evolved naturally without anybody understanding how they work. Functioning rockets have not.

Re:No

[Baron_Yam]

To be a little clearer - we have an evolution-provided brain model to copy, we did not have an evolution-provided rocket model to copy.

I can copy Kanji without understanding the meaning of what I'm copying... or even the basic rules of grammar. However, I'm not going to be able to write a new novel in Japanese.

Re:No

[iggymanz]

false, all branches of AI were done in the 1960s and before. There is nothing of significance that has been done for decades other than having more RAM and faster CPU.

Re:No

[religionofpeas]

Adding more hardware is a large part of the recent progress, yes.

Re:No

[phantomfive]

progress will continue by exploiting specialized hardware for this particular function

Is someone building specialized hardware for that? Last I checked it was still all just GPUs.

Re:No

[thinkwaitfast]

Moore's law is about transistor size, not computing power.

Re:No

[thinkwaitfast]

Or creating a new superconductor without understanding how superconductors work.

Re:No

[phantomfive]

Yeah, that's a good example.

Not So Fast

[Anonymous Coward]

I think a super smart AI is mostly a threat to the C level group, VP, EVP, etc. I always found it funny that in a big corporate environment most of the money goes to these people but yet are immune from any outsourcing campaigns.

http://dilbert.com/strip/2015-05-22

Eliza...

[__aaclcg7560]

The said the same thing about AI's back in the 1980's. Today we're no closer to having a self-aware AI program to replace psychologists.

Re:Eliza...

[religionofpeas]

Not yet psychologists, but we do have AIs that can analyze mammograms better than human doctors.

Re:Eliza...

[Shotgun]

And we also have AI's that can land passenger planes better than pilots (on clear days with smooth air). The commonality with mammogram analysis is not intelligence, but attentiveness. Analyzers get tired of looking at the same gray squiggly lines and miss stuff. Pilot's get lazy after a 10 mile stabilized approach that looks exactly it did the last 50 times she flew into the same airport.

Computers aren't smart, but they don't get lazy.

The Ignorance of Denial.

[geekmux]

(Claim in 1939) "We're going to put a man on the moon within 30 years."

(Society in 1939) "What a fucking moron. That will never happen."

Since the feedback here is essentially the same as it would have been in Society circa 1939, perhaps we should re-think the ignorance of automatically denying this claim, or dismissing Mr. Son as some kind of idiot. You would have been fucking wrong to do so in 1939.

What does the next 30 years bring? I'd say we can't even dream what may be possible. 30 years ago, the internet didn't exist. Chew on that fact for a minute.

Re:The Ignorance of Denial.

[Anonymous Coward]

30 years ago, the internet didn't exist.

Wrong.

Re:The Ignorance of Denial.

[religionofpeas]

I don't know if Mr Son is right or wrong, but he doesn't seem to be an expert on the matter, so his is just another opinion.

Re:The Ignorance of Denial.

[Anonymous Coward]

I suppose in 30 years I could film an AI in a sound studio and claim I invented AI.

Re:The Ignorance of Denial.

[thinkwaitfast]

30 years ago, the internet didn't exist

If you are so wrong about this, how can you be correct about your other assertions?

Re:The Ignorance of Denial.

[geekmux]

30 years ago, the internet didn't exist.

Wrong.

Give me a break. Hand a website today a modem and get back to me.

Re:The Ignorance of Denial.

[geekmux]

30 years ago, the internet didn't exist

If you are so wrong about this, how can you be correct about your other assertions?

I'm well aware of the history here, but to clarify my statement, the internet as we know it today did not exist 30 years ago.

ARPANET is so prehistoric it only offers a genetic marker of TCP/IP. BBS is hardly compared to what social media offers. And a user today can't even fathom how the hell the internet ever even worked via dial-up.

If anything, these facts reinforce my assertions, because no one was capable of predicting the internet of today in 1987.

Re: The Ignorance of Denial.

[Anonymous Coward]

Wrong again.

https://en.m.wikipedia.org/wiki/Neuromancer

Re:The Ignorance of Denial.

[eaglesrule]

From wiki for TOP500:

Since 1993, performance of the #1 ranked position has grown steadily in accord with Moore's law, doubling roughly every 14 months. As of November 2014, Tianhe-2 was fastest with an Rpeak[6] of 54.9024 PFLOPS, is over 419,102 times faster than the fastest system in November 1993, the Connection Machine CM-5/1024 (1024 cores) with Rpeak of 131.0 GFLOPS.[7]

And then there's this article from a company that knows a little bit about parallel processing.

As the trend continues for increases of processing power while cost decreases, the only real question to me is at what point is there critical mass enough where simulation can evolve strong AI. Unless there is a firmly held belief that our consciousness is the result of a spirit entity which is destined for a higher plane of existence after temporarily inhabiting a meat sack, why does it seem so difficult to assume that natural process that created human intelligence could not be reproduced and condensed into a fractional timespan, using software tools similar to Avida.

False

[Anonymous Coward]

They will replace humans in IRS queues for tax refund

62 Comments and no Mention of Skynet!

[sycodon]

You Young Whipper Snappers don't have a clue about evil A.I.!

Re:62 Comments and no Mention of Skynet!

[bob4u2c]

Skynet; now now sonny, let me tell you a story about your grandpa Colossus!

http://www.imdb.com/title/tt0064177/

Already Here

[JimSadler]

As long as a computer and its software is set up to solve specific problems they already exceed human abilities. Common, inexpensive chess software can destroy 99% of players and if one has a high power rig it will destroy any chess champion. Now imagine a computer set up to only seek rewards for opening new bank accounts of various sizes. A ten thousand dollar savings account may gain you a five hundred dollar bonus if you keep the account active for 90 days. If you turn that over four times a year you would get 20% in bonuses plus whatever interest the accounts earn in a year. And those accounts are insured so you have no risk compared to buying stocks and bonds. It goes without saying that you could have multiples of $10,000 to invest in such a scheme in banks all over the nation. One might also have a program that searches for distress sales for housing. Yes it would be wonderful to have one computer that can do it all and fit in your shirt pocket. But as far as functional intelligence computers already have humans beat. A few years back thee were AI programs that featured self learning from breeding other programs together and mating the best with the best over generations. Those programs became able to find quality solutions that humans could not understand the types of logic and rules by which the programs operated. Yet only these programs could find the solutions. In the New testament there is a remark that in the late days 'We will be confounded by our complexity.". i think we are already both confounded and enlightened by our complexity.

but humans ARE robots

[JustNiz]

Work 50+ years to enrich/empower the government, and are exactly told what to think by the media? puhlease we are already robots.

Re:but humans ARE robots

[Anonymous Coward]

"Robot" means slave in Czech, so yes.

Be careful generalizing

[sjbe]

IQ is like height in basketball. The best basketball players aren't the tallest people in the world but they are all taller than average.

A very good analogy.

two people with high IQ will out-perform a single person of super high-IQ.

That statement is task dependent. For some tasks it is true and for others not so much. There also are failure modes that multiple people are subject to that an individual is not. Much like your previous statement, crowds often are smarter than individuals but not universally so in all cases.

Also, there have been lots of data collected on IQs and success.

That is contingent on what you define as "success". I'm familiar with some of the studies you are probably referring to but be careful with such generalizations.

Re:Be careful generalizing

[Anonymous Coward]

There is really no reason to be careful about that statement.

Statistically there is no doubt about the fact that people who perform well in IQ tests are successful in almost every single field. However as several persons have pointed out before, high IQ in a person doesn't tell us all that much about that person.

lizardmen from the core (well their nanobots...)

[Anonymous Coward]

Maybe HPLDs avoid drawing the attention of dummos.

Super Smart CEO

[Anonymous Coward]

So how long till we have a super smart SoftBank CEO so we can all give a damn what he thinks ?

Re:Super Smart CEO

[Anonymous Coward]

I know you are just trolling but it's a valid point. Yes, why?
Capitalism is the vehicle of industry and the measuring stick of success. Anyone who has made their fortune using this system obviously has enough insight that you should give a damn about what they say. Or do you have better insight based on your proven track record of success?

There's been some miscommunication.

[Gravis Zero]

"Robots" is just what he calls the custodial crew. ;)

Not on my property

[Anonymous Coward]

I indent to stop any robot entering my property. I figure a 12g slug to the logic board will be sufficient in most cases. Electronic trespassers will not be tolerated.

Re:Not on my property

[Bob the Super Hamste]

You may want to upgrade to a full power rifle in the .308, .30-06, or 7.62x54r range as a 12 gauge slug lacks the penetration power. If that isn't enough there is always the Barrett line of firearms.

Re:Not on my property

[Anonymous Coward]

I indent to stop any robot entering my property.

Python won't help you here, man. Those droids are running Perl. -PCP

Re:Not on my property

[Anonymous Coward]

My property is small enough that the range of a 12g is sufficient, as I'm likely to simply walk 50 feet from my door to my perimeter and take out the offensive device point blank. And I live in a heavily populated urban area, so firing a high powered rifle is no bueno.

Re:Not on my property

[Bob the Super Hamste]

Range wouldn't be the concern but penetrating power. Slugs are broad heavy and slow so against unarmored things they work great but add some armor and a bit of distance between the armor and critical systems and a slug becomes ineffective. A high powered rifle on the other hand would make it through and still damage the circuits. Against unarmored devices I would probably suggest some finer buck shot like #0 so it is big enough to cause a lot of damage but you still get a spread to increase you chance of hitting critical areas.

And it still won't be considered A.I.

[Maxo-Texas]

The bar will keep being moved. Even when computer programs are 'smarter' than 99% of humans, they won't be considered "real A.I." by many people. Some for religious reasons. Some for property reasons.

It may be not one versitile A.I. that can do everything better than humans- but hundreds of "idiot savant" who each excel human beings in a particular area.

circular logic 101

[epine]

A computer will have the IQ equal to 1,000 times the average human by that point, he said.

If the bar is the IQ of someone who presently thinks this is the way IQ works, he's probably right.

AI can't split compound sentences.

[Anonymous Coward]

AI is a useful tool, but an utter joke. It can't even understand language. We keep adding tricks, but they still have no idea what is going on.

Sort of confused at what you are shooting for...

[TiggertheMad]

So, you have gone back and confirmed that I have consistently said that AI is a tough field, and that we are a lot farther out from generalized AIs than CEOs would have you believe. Check.

Then you pull in a quote from a conversation with a ms game dev. I am not quite sure what point you were hoping to illustrate with that. The point I was making was that game devs of the time weren't even trying to build a intelligent, learning system that would adapt to player behavior or environmental changes, but they simply took the lazy/easy path of just peeking at player input and using asymmetrical information to appear to be smarter than they actually were. I was appalled at the intellectual laziness of the dev, because they hadn't even attempted to experiment with more nuanced approaches but simply waved them off because it was simpler to just let the AI cheat.

Now, I get that there are real world time and budget constraints, especially with game dev work. However, the attitude I encountered was akin to 'I think PI = 3, because anything more is too much work, and setting PI = 3 has worked on all the bridges I have built so far'

I am a little confused though, on how either of these points leads you to the conclusion that 'Academic Techniques' aren't adequate for real world problems. Some of the best and most exiting work in the 'real world' being done by big companies is built solidly on academic techniques. Go read about Google's machine translate work, for example. It is built on a neural net model, and is making some pretty amazing progress.

Finally, if you hope that using my own opinions about the state of AI will somehow shore up your opinion of academic AI techniques, I will be the first to claim that I am a talented amateur at best. Build your arguments on my thoughs on the topic, and you are truly building a house on sand!

CEOs and other Imbeciles < My opinions < Real Researchers who know their shit

Money != Smart

[TiggertheMad]

Yep. That guy. He is an idiot, and every time I see him open his mouth, my opinion falls further. He knows business, and has got some pretty cool electric cars made, but he keeps saying bat shit crazy things about tech fields that I don't see him as technically qualified to discuss authoritatively. I would love it to have some brilliant Tony Stark style billionaires running around in the world, but having a lot of financial success doesn't meant that you are good at solving any problem other than making a lot of money.

BWT, signing checks for AI research does not equate to being a Fuzzy Zadeh, John Holland, etc.

Re:Money != Smart

[mbkennel]

Musk is a firm reality-acceptor when it comes to the technological issues facing his businesses: energy storage, battery manufacturing, car manufacturing, and liquid kerosene fueled rockets. He knows how hard they are and what it takes.

If a software company CEO said oh 'look at the progress in batteries, in 30 years, they will have 50 times the energy density of gasoline', he'd honestly be upset by this wild overselling, and the scientists will describe the energy and chemical constraints of the laws of physics that make that prediction ridiculously wrong.

But he's doing the same thing with AI.

Re:Money != Smart

[khallow]

But he's doing the same thing with AI.

There's much more room for improvement with AI than there is with energy storage via chemical batteries. The latter is strongly constrained as you noted:

and the scientists will describe the energy and chemical constraints of the laws of physics that make that prediction ridiculously wrong

Where's the corresponding constraints on AI? We're nowhere near most of the physical constraints on computation. More relevantly, our computer and software systems have experienced enormous bloat over the decades of Moore's law. Getting the same results out of a factor of 50 fewer CPU cycles is not an unreasonable expectation. A strong AI might be able to get a lot more work than that out of the same number of CPU cycles.

"Come w/ me if you wanna live..."

[Anonymous Coward]

"I'm here to help you - I'm Reese Sgt. TechCom DN38416 assigned to protect you" via APK Hosts File Engine 9.0++ SR-7 32/64-bit https://www.google.com/search?hl=en&source=hp&biw=&bih=&q=%22APK+Hosts+File+Engine%22+and+%22start64%22&btnG=Google+Search&gbv=1/ "you've been TARGETTED for TERMINATION!"

Best shield vs. online threats & speed gains (others slow you). Does more 4 less vs. ANY 1 illogical "Bolt on 'MoAr' so-called exploitable 'security solution'" (DNS/antivirus) using more doing less (dns/antivirus/browser addons)

As "that terminator is out there: It can't be bargained with. It can't be reasoned with. It doesn't feel pity, remorse, or fear & it absolutely will not stop EVER (until U R DEAD)"

P.S.=> "It's a HYPER-ALLOY Combat Chassis - Microprocessor controlled: FULLY armored, VERY tough!"

Code's VIRUS-PROOF (every proc/function checks vs. alteration) & crash-proof (via std. structured errhandler refactored override reset on abend)

Re:"Come w/ me if you wanna live..."

[Anonymous Coward]

Your script is useless and obsolete.

CORRECTION!

[downright]

Supersmart *SEX* Robots Will Outnumber Humans

History in the making....

[WolfgangVL]

October 21, 2051, the ratio of synthetic to organic beings in the united states stood at nearly 10:1, in other less developed countries, it stood closer to 25:1. Fearing a repeat of the robotic riots of 2038, (which very nearly destroyed the entirety of the recently annexed New California Republic) the newly reformed United Nations ratified the doomed "Humanity First Treaty", which aimed to place the basic needs of organic humans above those of our synthetic counter-parts.

Within days, robotic and synthetic beings found themselves evicted from the tiny living spaces they had occupied for decades, in favor of humanities ballooning population. Their property deeds and rental leases found invalid under the new treaty, the developed words massive synthetic populations marched on the major population centers of the USA, New China, and Neo-London, simply because they had nowhere else to go. Having no need to rest, and access to real-time infrastructure data by their nature, these mechanical masses leveraged the near 100% up-time, and energy efficiency humanity had engineered into them to quickly organize themselves, and disrupt the complicated economic systems and massive supply chains that served to provide food and water to incredible numbers of these population centers. Within a week of the start of this worldwide robotic protests, starving humans in Auburn, WA, USA, fired the opening shots (on peaceful robotic protesters) in what today is known as "humanities greatest blunder", but at the time, was simply dubbed, "Decommissioning" or "The De-Con"

Time to refresh your memory

[Anonymous Coward]

The outcome was predicted many years ago by the Wachowski Brothers. Worth watching again as we approach that reality. http://www.thematrix101.com/animatrix/renaissance.php

There is one slight problem.

[pjv936]

It takes real intelligence to create really good AI and we ain't got it.

Smarter than CEOs

[wasteoid]

They will be / already are smarter than CEOs, but I doubt we'll see cushy CEO jobs disappear.

Re: The banality of ubiquitious genius will doom u

[wanfuse123]

When AI reaches human intelligence, it wont take years, months, or days to leave us in the dust, but seconds. It will update and improve its own code by millions of times almost instantly. It will process what humans would process in a million years in a matter of seconds and strategies its next million moves without us even knowing whats going on. Wise men fear to tread, where apparently humans don't. Fear the unknown! When it comes it will be in the blink of an AI, and to late to go back. All things that came before it will be meaningless.

Get ready for a surprise....

[Anonymous Coward]

Title of story: Supersmart Robots Will Outnumber Humans Within 30 Years

That's it folks, soon it will be time for the "un-elite" to "get out of the pool" and get your asses to Mars.

/.ers disagree #1/2... apk

[Anonymous Coward]

I'm going to continue using the Host File Engine. Your software is well written, functional. The Host File Engine performs exactly as promised by mmell

his hosts program is actually pretty good by xenotransplant

his hosts tool is actually useful for those cases in which one does indeed want to locally block stuff outright while consuming minimum system resources by alexgieg

I've never tried to belittle (APK's) work, I've flat out said it's good by BronsCon

take a look at the APK hosts file engine by SuperKendall

APK is kinda right. I've tried his hosts file generating software. It works by bmo

APK is totally right on this count. Adblock Plus on Firefox mobile is a dog on older, or lower end, phones. A hostfile based adblocker makes for a much better experience by chihowa

I like your host file system by Karmashock

I find your hosts file admirable by vel-ex-tech

* My code's liked + recommended & hosted by Malwarebytes' hpHosts!

APK

P.S.=> More coming... apk

/.ers disagree #2/2... apk

[Anonymous Coward]

I support APK's stand on the hosts file by Trax3001BBS

Your premise that hostfiles are a good way to deal with advertising and malvertising is quite valid by JazzLad

No complaints from me, I like APK... Reminds me to use a host file. Also, his stuff is free by aaaaaaargh!

APK's monolithic hosts file is looking pretty good by Culture20

APK... Awesome to see he's still spreading the good word by Molochi

ABP is insufficient as a solid hosts file does everything that APK reminds us about by fast turtle

APK isn't wrong by cfalcon

APK, I know people give you a lot of shit regarding hosts, but please don't ever stop by nasredin

You need APK's hosts file by Teun

APK solution STILL relevant by Thud457

you're right about hosts files by drinkypoo

APK

P.S.=> They're in addition to https://hardware.slashdot.org/comments.pl?sid=10300111&cid=53942071/ many more earlier + 1,000's worldwide - there's no arguing w/ success... apk

Security & web experts disagree

[Anonymous Coward]

Oliver Day (SYMANTEC/SECURITYFOCUS):

http://www.securityfocus.com/c...

"The host file on my day-to-day laptop is now over 16,000 lines long. Accessing the Internet -- particularly browsing the Web -- is actually faster now."

"... More recently, projects like Spybot Search and Destroy offer lists of known malicious servers to add a layer of defense against trojans and other forms of malware"

OReilly on hosts for security -> http://oreilly.com/pub/a/windo... & For speed -> http://www.oreillynet.com/pub/...

Steve Gibson endorses hosts as good https://www.grc.com/sn/sn-045....

Aryeh Goretsky of ESET/NOD32: hosts = good security http://it.slashdot.org/comment...

Brocke Wilders of WILDERS' SECURITY does via an inferior clone of MY PROGRAM http://www.wilderssecurity.com...

Malwarebytes hpHosts' Admin hosts + RECOMMENDS my work http://hosts-file.net/?s=Downl...

APK

P.S.=> Eat your words... apk

Re:Security & web experts disagree

[Anonymous Coward]

You brag about your little script because it's the closest thing to a functional program you've ever managed to write.

You shriek with impotent rage at anyone who criticizes your little script because you're insecure about your lack of technical skills.

You will now prove me right again.

AI is not "computers"

[cjonslashdot]

Today's AI systems are not computer programs: they are neural networks. Many AI systems use computers to simulate neural networks, since neural network hardware is hard to come by, but the underlying model is not a von Neumann computing model - it is a neural network model.

We already have disposable sentient automatons...

[Anonymous Coward]

......they're called Indians and Asians.

Re:Sort of confused at what you are shooting for..

[khallow]

The point I was making was that game devs of the time weren't even trying to build a intelligent, learning system that would adapt to player behavior or environmental changes, but they simply took the lazy/easy path of just peeking at player input and using asymmetrical information to appear to be smarter than they actually were.

In other words, when you slightly change the rules about how AI is supposed to work, the problems turned out so easy that the developer didn't need to bother with any formal AI approaches.

It's also worth noting that the developer solved the problem. Excessive problem description and feature creation is a notorious killer of many academic projects not just in the AI world. The business world occasionally falls prey to that as well, but as we see here, not always.

I am a little confused though, on how either of these points leads you to the conclusion that 'Academic Techniques' aren't adequate for real world problems. Some of the best and most exiting work in the 'real world' being done by big companies is built solidly on academic techniques. Go read about Google's machine translate work, for example. It is built on a neural net model, and is making some pretty amazing progress.

First, on your machine translation example, "amazing progress" compared to what? Both neural nets and machine translation have been around for decades. The "wow" factor of Google's efforts comes from the infrastructure that has been built up (being able to copy/paste something something to be translated over the internet effortlessly and throw orders of magnitude more CPU cycles at it) rather than the algorithm.

What I consider a more relevant case of doing something new with neural networks is Google's Deep Dream where one uses a neural network trained on finding certain images (say like images of buildings) to iteratively perturb images (like a mundane landscape photo) to bring out those patterns (ending up with a weird, psychedelic image with lots of buildings crammed into every part of the image).

Unfortunately, there's not a lot of academic precedent for that. The related research articles heavily emphasize classification and detection improvements not the wow of turning a boring image into piles of buildings or whatever. Going to the games genre, this would be an excellent way for a neural network to create on the fly themed maps and art for a game. Train a neural net to spot the desired sort of maps or artwork and then starting with a sufficiently simulating pile of mush, bring out the desired patterns iteratively in the mush.

Finally, if you hope that using my own opinions about the state of AI will somehow shore up your opinion of academic AI techniques, I will be the first to claim that I am a talented amateur at best. Build your arguments on my thoughs on the topic, and you are truly building a house on sand!

You made the claim that academics are at least a century out from building anything resembling human or higher level AI. That says right there that you don't think they have much to say about the subject now. This brings up my second point, Your beliefs are inconsistent. We don't need to care about any validation of my beliefs when the conflicts in your beliefs are more than ample to defeat your assertions.

The most obvious source of any AI development is completely missed here. It's not academics, CEOs, or secretive government agencies. It's computers. Once you've completely automated the creation of human or better level AI, then it's not going to need a century to get there. You might not even need a day.

Bootstrapping more sophisticated algorithms from existing one that have sufficient power to improve themselves is the great missing step here, I think. And modern AI research simply isn't going that way at present. I think at some point that will change, then we'll come up with more relevant concerns than how many more centuries we'll wait till humanity does this thing.

Don't forget us.

[Anonymous Coward]

"I believe this artificial intelligence is going to be our partner," he said. "If we misuse it, it will be a risk. If we use it right, it can be our partner." ...and if it replaces the bulk of the tech workforce, it will be a risk. Think an entirely new demographic added to your nation's poverty level. Numbering in the millions.

yeah,i heard that

[Anonymous Coward]

already 30 years ago

Optimization

[manu0601]

A smart AI could ask itself what human are good for, and whether their presence is really useful. Next step would be to "optimize" humanity so that it consumes less resources. A few specimen could remain to preserve biodiversity.

Re:Optimization

[RespekMyAthorati]

A smart AI could ask itself what human are good for

Absolutely nothin',
Huh!

FAR from it per other examples

[Anonymous Coward]

Windows NT Magazine April 1997 "BACK OFFICE PERFORMANCE" pg 61

(For SuperSpeed.com PAID CONTRACT (wrote parts of SuperCache 40% performance boost) & SuperDisk finalist @ MS Tech Ed 2x in a row 2000-2002 HARDEST CATEGORY: SQLServer Performance Enhancement)

WINDOWS MAGAZINE 1997 "Top Freeware & Shareware of the Year" issue pg 210 #1 entry

PC-WELT FEB 1998 pg 84

WINDOWS MAGAZINE, WINTER 1998 pg 92 MUST HAVE WARES

PC-WELT FEB 1999 - pg 83

CHIP Magazine 7/99 - pg 100

GERMAN PC BOOK Data Becker "PC Aufrusten und Repairen" 2000

HOT SHAREWARE #46 issue pg. 54 (PC ware mag from Spain) 2001

Paid for article @ PCPitstop in 2008 http://pcpitstop.com/news/winn...

UltraDefrag64 Process Priority Control credited by lead dev -> http://ultradefrag.sourceforge... or here http://sourceforge.net/tracker...

APK

P.S.=> Topoed off w/ 40 enterprise class systems 1994-2008 that run entire companies... apk

Re:FAR from it per other examples

[Anonymous Coward]

Exactly. You've done nothing. There's no other reason you'd brag about a forum post you made 10 years ago, or a useless little script you wrote nearly 20 years ago.

You have proven me right as I said you would.

You will now prove me right again.

Good. No more immigrants needed.

[Anonymous Coward]

No more "immigrant" aliens, illegal or otherwise needed. Sorry Haji, but the boat is full.

Re:Sort of confused at what you are shooting for..

[AK Marc]

Game AI can beat a human every time because the parameters are sufficiently narrow, and a computer can calculate faster than a person. That's not AI, that's brute force to look smart while being dumb.

The problem is that AI's definition is essentially "anything that seems smart" and under that definition, bad game AI fits. Deep Learning is not AI, and has no path to strong AI. Most current AI research has no path to strong AI. Academic AI has almost given up on strong AI, and has taken a detour in demonstrating weak AI on stronger computers that looks like stronger AI, when the AI is no better, but the weak AI is faster.

CEOs and other Imbeciles < My opinions < Real Researchers who know their shit

And you put everyone with an opinion you don't like into the "imbeciles" group. By reasonable measure, that puts you in the imbecile group.

Re: The banality of ubiquitious genius will doom u

[ChrisMaple]

You are making the assumption, neither proved nor demonstrated, that there will be a factor-of-one-million improvement of code possible. You are also assuming that we'll allow a machine to alter its own code.

WTF? LMAO - learn to read skulking worm!

[Anonymous Coward]

See my subject unidentifiable little hiding "ne'er-do-well" slime & my posts all made you eat your fav. dish: YOUR words, lol!

* Projecting that YOU are a do-nothing "JEALOUS JOWIE" too?

APK

P.S.=> Utterly priceless - but that's the PRICE of your WASTED LIFE - forced to be a miserable little SWINE skulking in the shadows KNOWING you're a fatherless bastard (you must be raised by a woman only because you act like the bitch you are, hahahaha)... apk

The rise of the dysgenic masses

[Anonymous Coward]

This will lead (even further) to the rise and rise of more and more stupid people, who can't function without leeching off the hardworking, intelligent people in our societies, until they take over and are unable to repair the robots that give them the very luxury they think is their birthright.

Most human beings are completely incapable of basic rational thought - just ask anybody you know if they think white people have the right to have their own countries, see how long it takes them to reply, while their cognitive dissonance kicks in...

Clear propaganda at this stage...

[Anonymous Coward]

These "robots will make us obsolete!" stories are such clear propaganda now - Slashdot needs to stop being a constant fúcking voice for this kind of propaganda...same with constantly pushing the Universal Basic Income (which is always paired with the AI fearmongering), which is a plan chock full of insurmountable flaws, that requires completely ignoring how easily corrupted the the UBI can become, through political influence (hint: there's a very fine line between a utopian functional UBI, and a fúcking dystopia caused by the political weaponization of the UBI, to destroy the welfare system and progressive taxation system...).

Cut it out with the bullshít propaganda. Get back to tech news - not tech industry propaganda.

Yawn

[ebvwfbw]

I remember in the 1980s when AI was supposed to take over the world. As a computer scientists my job could be eliminated in JUST 5 YEARS! Why even bother? The horror. I said - yea, yea... I'll believe it when I see it.

Same old crap, just new people saying it. I don't see anything new that worries me in the AI area.