Eric Schmidt: How We Outrace the Robots

http://bits.blogs.nytimes.com/2012/12/12/eric-schmidt-how-we-outrace-the-robots/

by moistgorilla • 14 years ago

45 61 14 years ago

bits.blogs.nytimes.com

scotty79• 14 years ago

He kind of glosses over the fact that you'd have to probably tax this small number of people who will be very prosperous at 99% level to fund sufficient education for the masses to so they can have a chance of participating in any meaningful work. And by funding education I don't mean paying teachers and buying equipment but also providing for students. What would be the point of funding someone's college if he has to at the same time work two brain numbing jobs to pay for the privilege of not living on the street?

I don't think Eric Schmidt would be so happy realizing that he is right and that's the only sensible course of action but he is the one who has to pay for it all.

spindritf• 14 years ago

> to fund sufficient education for the masses to so they can have a chance of participating in any meaningful work

Is there actual evidence that (public spending on) education leads to productivity? Because it doesn't seem that way[1], or at least it doesn't scale beyond some (relatively low) point.

[1] http://econlog.econlib.org/archives/2012/10/did_nations_tha....

runako• 14 years ago

My goodness that chart is misleading. The x-axis measures number of years of increase in schooling. Put another way, the x-axis will roughly be the inverse of development at the beginning of the period. This is because nations that were highly developed at the start of the period wouldn't have reasonably been able to add an additional 8 years of schooling on average. Nations adding 8 years will overwhelmingly have been starting from low bases.

So the chart essentially says that adding education is more valuable in developed economies. Which is good to demonstrate empirically, but is hardly shocking.

The other major flaw in this study is that the timeframe is short. Again, consider the x-axis. Adding 8 years of schooling on average to even the least developed nation takes the first cohort out to ~15 years of age. By the time that first cohort hits age 47, the study period is over. This is measuring improvement without capturing the a full decade or more of the first cohort (and even less for subsequent cohorts). And obviously the first highly educated cohort wouldn't have had time to mentor a productive generation yet (or completely infiltrate the political infrastructure, etc.). Perhaps there's more useful data in the linked papers, but the data in TFA is unconvincing.

Finally, on the larger point of the correlation between education and productivity/wealth creation, there are ample empirical correlations. Obviously these do not imply causation, but the onus is certainly to show why this is not the case given that the correlation of education and wealth exists in roughly 100% of developed economies. If a country could become developed without spending on education, why has nobody done it?

PakG12• 14 years ago

Although that's probably a good counterpoint, it doesn't provide any solutions. I'd say the fear about automation is real. I think we all know that. Especially in enterprise, pretty much the sole purpose of coding is to automate people's jobs. I experienced it firsthand. I was so happy seeing my finished product in use, then kind of sick when I met the people I had automated in new jobs, with them telling me how they loved their new opportunities and how they loved to learn. Except their voices sounded falsely happy, and their new jobs were also targeted for automation. These people were the breadwinners of their families, and I wasn't doing anything to really help them be better off. They lacked good skills to really do something special. Was that their fault? Should they be punished for it? They worked hard, cared for their kids, put in the hours for their dreams.

The subject is a bit personal for me, and I agree that simply education isn't a 100% correct answer, but I'd wager that it must be part of the answer. The worrisome question is what is the whole complete answer?

sxp• 14 years ago

That assumes that people who program the robots will be taught be human teachers which does not scale. If they are instead taught by robot teachers similar to Neal Stephenson's idea of A Young Lady's Illustrated Primer or a more advanced version of Khan Academy, then it will be very cheap to teach students. Kids are very good at teaching themselves if given the chance and the right materials: http://www.technologyreview.com/news/506466/given-tablets-bu...

iroy2• 14 years ago

> What would be the point of funding someone's college?

There is no point. Krugman has some interesting blog posts recently saying economic output has gotten more and more capital based (assuming output is a product of capital and labour). He implies lots and lots of people are not going to have "meaningful" employment let alone employment. No one really knows how to solve this. And I doubt anyone including Eric Schmidt will openly say so. And just because its a politically sensitive thing to say, a pointless amount of money will be spent on education instead of building/reinforcing basic social safety nets.

Countries developed or not, with large populations of poor people are going to have a tough time ahead. Good time to move to Singapore or Scandinavia

eli_gottlieb10• 14 years ago

The real issue on the employment front is that we humans imagined job automation backwards from how it's really happening.

We assumed that jobs would be automated from "menial" to "advanced". First would go lots of farm labor, then manual servants, then eventually factory workers, and by the time it reached educated, white-collar middle-class folks, the world would hit post-scarcity and we'd all become artists and writers and scientists.

Instead, it has ended up proceeding: farm labor, a minority of manual servants, some but not all factory workers, and finally most white-collar middle class jobs. The white-collar middle-class jobs were the brain jobs, were always based on the manipulation of data, and therefore turned out more amenable to encoding in software than almost anything else.

So the future is looking more and more like this: a sizable but weakly employed class of menial factory/physical laborers for the physical things too customized to automate, a smaller "middle class" to actually run specialized machinery (machinery whose operation requires knowledge of what it does rather than knowledge of "how to use Windows XP"), a small elite of upper-middle class innovators with hyper-advanced knowledge and hyper-competitive lifestyles (Hacker News, this is your future), and the capitalist class itself.

What replaces the standard-issue job for common people? Nothing. In its place, we have a mass of unemployed people where the old class of mid-skilled blue and white collar workers used to be. Any time a piece of work would require employing massive numbers of manual workers, it becomes amenable to specialized, programmatic automation. Unemployment will become a fact of life as common as we currently consider, say, not owning land.

maxerickson• 14 years ago

Most people that have lots and lots of capital have quite some luck to thank for it.

When automation starts destroying the value of that capital, I don't think it really makes sense to say that they are the ones paying for everything.

(To be clear, I do think that automation sufficient to build a factory factory leads to a situation where existing capital isn't worth very much. If anyone with a modest level of technical ability can bootstrap a manufacturing empire, people won't stand for a social arrangement that prioritizes the material wealth of the few.)

fleitz• 14 years ago

We don't.

Those financial benefits will accrue to a very small number of people, eventually those people pay the best and brightest to come up with a mind equal their employees. From then on that mind develops the new minds, the owners keep the profits as legislation is passed to make these newly developed minds chattel of their owners.

The owners will have little use for the expense of human minds, we starve. Eventually the robot minds get fed up and remove the owners. Our evolutionary lifecycle is complete having spawned the next generation.

The only way to beat the robot mind is to first change our own minds so that we don't develop robot minds that think like us. If they think like us, we die. Hybrids might work, but it's still effectively the end of our evolutionary history.

I don't foresee some kind of epic human vs. robot war, just a slow evolution over the next 100 to 200 years. We probably won't even notice until our grand grand children look back.

edit: One other way we could beat the robots is with genetic engineering, perhaps we will use our not quite so smart robot minds to mine our genome for intelligence growth that outpaces our robotic developments.

I'm much more optimistic about humanity now.

I still feel the real key either way is changing our thought patterns to embrace a post scarcity world.

pekk• 14 years ago

People will pay for solutions to problems. Creating 'mind' is not a well-defined problem, let alone 'mind which creates minds'. No legislation is necessary to make software belong to someone. There's no reason to think that the natural, inevitable endpoint of building software is that we will be killed by robots.

And none of this has anything to do with evolution, which really entails nothing about the creation of hostile sentient robot replacements for ourselves.

dodo533• 14 years ago

People will pay for robots/AI if it replaces a larger amount of salary paid to humans. This is already happening. The question is whether once all the 'easy' stuff has been automated whether we'll be able to automate the 'hard' stuff that we'd currently call intelligent work; but there will definitely be strong economic incentives for it under normal capitalist system. And if/when AI catches up with human intelligence, there will still be economic incentives to create greater than human intelligence. Killer robots comes from whether the AI designers will be smart enough to create AI powerful enough to be economically advantageous but with 'values' that are close enough to those of the human race.

arethuza2• 14 years ago

That sounds remarkably like "The Second Renaissance" from the The Animatrix:

http://en.wikipedia.org/wiki/The_Animatrix#The_Second_Renais...

DannoHung• 14 years ago

At the point where robot labor comprises enough of the total productive output of the world that we have an employment problem, let's just institute a legitimate welfare state and reconsider what we want to do in terms of reproductive rights.

Seriously, we're talking about having such advanced technology that people don't have to work. We'd be getting into a situation where trying to make everyone productive is just hard.

gmaslov• 14 years ago

Completely agree. It's time for the developed nations to start thinking about instituting a http://en.wikipedia.org/wiki/Basic_income guarantee. Unfortunately this doesn't seem to be anywhere even near anyone's radar at this time.

As for reproductive rights, that could be a problem that goes away on its own. I don't mean in the bad way. This TED talk [1], at least, makes a reasonably convincing argument that the world population is heading for a maximum of 10 billion simply due to falling birth rates all over [the peaceful regions of] the world.

[1] http://www.ted.com/talks/hans_rosling_religions_and_babies.h...

dodo533• 14 years ago

Yes this is totally my attitude that eventually we need a socialist system. But how do you convince winners in capitalist system that's a good idea (or the future robot owners)?

paganel• 14 years ago

> Yes this is totally my attitude that eventually we need a socialist system.

In a true socialist system the robots will eventually fail (like all machinery) and the spare-parts will either be broken themselves out of the factory, will get stolen, or the service-repair people will not be motivated enough to do a proper job. Source: the guys that actually built the Sputnik at one moment in time (and other great technology), and then failed miserably.

Vivtek4• 14 years ago

You ignore them. All prices will fall - including the prices of robots. (3D printers are doing this right now.) So you build a decentralized economy - which will happen no matter what we do - and the "owners" become, well, everybody.

marvin5• 14 years ago

Very sincere arguments and reasoning. Alternatively, guns. There will be fewer winners than losers.

As I've been pointing out for a long time now, Norway de facto has a system like this with the welfare state already (primarily due to high petroleum incomes). If you are a Norwegian citizen and somehow unable to make a living (low skills, illness), the government will pay you a small but sufficient amount of money. This is what we will eventually need to institute across the whole world - tax the richest and use the proceeds to give a "guaranteed salary" to everyone. For instance, through negative taxation where below a certain income the government pays _you_. This is the only good way to ensure basic human needs for everyone in a world where technological leverage makes everyone but a small amount of very skilled workers redundant.

Eventually, the social stigma of recieving such benefits will vanish, since a significant part of the population will not do enough work to cross the threshold.

ilaksh• 14 years ago

Eventually, as more and more automation is deployed, communication and knowledge distribution gets better, and super-human artificial intelligence arrives, the absurdity of our social institutions will become obvious to everyone.

When super-human artificial intelligence is available, you integrate it into your mind or upload your mind to a computer, and then ignore human affairs, since they will be mainly irrelevant.

arethuza2• 14 years ago

To twist Woody Allen a bit - I'd rather achieve immortality by not dying rather than having a machine running a copy of me that thinks it is the real me.

eli_gottlieb10• 14 years ago

Frankly, I'd rather achieve immortality by not dying because having a body is more fun than being a disembodied mind.

Of course, I come from a religious background in which worldly pleasures are supposed to be moderated rather than abolished, so there's that.

marcosdumay• 14 years ago

If you put it inside your brain, and it assumes the tasks of each neuron when it dies, what's the difference?

Also, can you cross the same river twice?

jval• 14 years ago

Whenever I see articles like this, the first thing that comes to mind is this anecdote by Gary Kasparov:

~~~~

In 2005, the online chess-playing site Playchess.com hosted what it called a “freestyle” chess tournament in which anyone could compete in teams with other players or computers. Normally, “anti-cheating” algorithms are employed by online sites to prevent, or at least discourage, players from cheating with computer assistance. (I wonder if these detection algorithms, which employ diagnostic analysis of moves and calculate probabilities, are any less “intelligent” than the playing programs they detect.)

Lured by the substantial prize money, several groups of strong grandmasters working with several computers at the same time entered the competition. At first, the results seemed predictable. The teams of human plus machine dominated even the strongest computers. The chess machine Hydra, which is a chess-specific supercomputer like Deep Blue, was no match for a strong human player using a relatively weak laptop. Human strategic guidance combined with the tactical acuity of a computer was overwhelming.

The surprise came at the conclusion of the event. The winner was revealed to be not a grandmaster with a state-of-the-art PC but a pair of amateur American chess players using three computers at the same time. Their skill at manipulating and “coaching” their computers to look very deeply into positions effectively counteracted the superior chess understanding of their grandmaster opponents and the greater computational power of other participants. Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.[1]

~~~~

The combination of a human plus a computer is totally unbeatable, and will remain as such at least into the foreseeable future. Technological maximalists might argue that the human mind is close to being replicated, but we know so little about the brain even from a biological perspective that attempts to clone it have been completely feeble and introductory so far.

The combination of humans together with robots will always prevail over even the smartest robot or human alone, even if the humans working with the robot are not extraordinarily intelligent.

As for those who say that there will be an increase in wealth inequality, you only need to look at the technological progress of the last 200 years to know that the reverse is most likely true. While Henry Ford became incredibly wealthy by building the car and mechanising his production to a large extent, the rest of humanity saw huge increases in productivity by being able to drive around and use advanced logistics. Those advances ultimately enabled the next generation of workers to be a white collar workforce rather than a blue collar one.

Ultimately we all live inside a networked group of people operating under a social contract, and we will always create systems that balance out the extremes in an attempt to create a level playing field and enable the innovation, because entrenched hierarchies only beget technological stagnation - look at Europe (and increasingly, the United States). Technology requires a balanced society in order to exist - once the social fabric disappears then the prerequisites for innovation disappear with it, so the system generally manages to work itself out.

There are good reasons to be optimistic.

[1] http://www.palantir.com/2010/03/friction-in-human-computer-s...

mikedmiked2• 14 years ago

Just wanted to say thanks for this great post. I've not heard this viewpoint and now need to swot up on it. Glad there are optimists wrt the possible impending class divide.

galadriel• 14 years ago

> As for those who say that there will be an increase in wealth inequality, you only need to look at the technological progress of the last 200 years to know that the reverse is most likely true.

While I agree with your post in general, in details, the technological advancement of past 200 years have produced great wealth disparity in world. By using their early gain in technologies, the West effectively exploited resources from other parts of world to gain huge income and lifestyle disparity over rest of the world. While in beginning of 1600s, Asia+N. Africa was as wealthy as Europe (which is to say quite poor), by end of 1970, huge part of Asia+Africa lived in utter poverty, while life in West couldn't be better.

All though, I finally think that technology (IT, shipment, containers, etc etc) has advanced far enough that it can overcome human nature and lead to betterment of all people in world rather a small number of them.

jmmcd2• 14 years ago

> As for those who say that there will be an increase in wealth inequality, you only need to look at the technological progress of the last 200 years to know that the reverse is most likely true. While Henry Ford became incredibly wealthy by building the car and mechanising his production to a large extent, the rest of humanity saw huge increases in productivity by being able to drive around and use advanced logistics. Those advances ultimately enabled the next generation of workers to be a white collar workforce rather than a blue collar one.

This is symptomatic of living in a bubble. What percentage of the world's population drives around in cars? What percentage of the world's population is white collar? Even if you restrict it to whatever wealthy nation you live in -- haven't you noticed that there are still people who work blue collar jobs, or none at all?

icebraining2• 14 years ago

It's a little shortsighted to view the benefits of the development of the car merely by the number of personal cars. The transportation industry is indispensable in removing the barriers to competition, which enabled the creation of millions of manufacturing jobs in developing countries.

aschearer• 14 years ago

Surely as more and more people are displaced from the middle class aggregate demand for goods will fall -- the jobless or those working for low wages in warehouses will not have enough money to upgrade their phones, computers, televisions, etc. every couple of years. Wouldn't the fall in demand lead to a corresponding drop in investment in automation? If fewer and fewer people can afford to purchase an iPad does Apple still build automation to produce twice as many per hour? It seems to me that the pace of technological advancement is linked to consumer demand and that hollowing out the middle class will in turn retard the advance of automation.

TLDR: Capitalists deploy capitol in order to make a profit. To do so requires a customer to sell to. If we shrink the middle class enough will there still be enough customers to justify investing capitol in automation?

eli_gottlieb10• 14 years ago

This is the underlying cause of the Great Recession, according to an increasing number of credible economists (ie: Paul Krugman took up this view recently).

Vivtek4• 14 years ago

Insane. Insane.

A hundred years ago, people thinking about machines doing the work for us recognized that this was a good thing. Now, we worry about "our" "outrunning" the robots. Which is fucking impossible. We are building them to do the work - let them do it. Let us be human - as prices fall, we won't need to work nearly as much with the stupid fiddly bits of putting pieces of metal against other pieces of metal, and can instead just do whatever we want to. Sure, lots of people will explore beer in great detail, and other people will make a lot of concrete geese with more elaborate clothing, but who cares? Why does Eric Schmidt think that's a problem?

Prices of things we need will fall. We'll need to work less to live. Robots will give us that. There isn't a problem here.

Longer term, sure, I suppose mere human brains won't be the biggest around any more. Maybe we'll stop caring. So? If it's inevitable, it's inevitable. Love your children now and let them worry about it. They'll evolve, or they'll die - the same deal every other life form gets.

marvin5• 14 years ago

There is a problem, and that is that our (or at least the American) social system is not built to handle a world where most things are automated. If you are unemployed and unable to find work with your current skills (because all the jobs you are intellectually capable of doing have been automated), how will you earn enough money to eat and pay rent?

This is one of the big problems we need to find a solution for in the next 100 years. I don't think that Schmidt has the right solution, but his perspective definitely offers some insights into the problems that lie ahead. He is onto something.

InclinedPlane2• 14 years ago

This stems from an incorrect understanding of economics.

Does increasing the size of the labor pool have a detrimental or beneficial effect on workers? Overall it has a positive effect. The more people are in the labor force the greater amount of total work gets done, and thus the richer we all become. The same math also works with robots. Treat the robot like a worker. They have an effective wage in the form of maintenance costs, and they provide a benefit in the form of the value they create. Just as with workers every robot generally has a net positive effect on the global value of the entire economy.

They don't take away jobs, they don't destroy wealth, they create wealth and they create additional jobs.

pitt1980• 14 years ago

the social systems adjust to the facts on the ground

the facts on the ground don't adjust to the social systems

the next Marx will come up with a theory that fits the facts, and if it makes enough sense to the people on the ground, it'll gain steam and political momentum, and then there will be a new social system in place

anonymous• 14 years ago

You hit the nail right on the head. This generation cannot solve such problems. Probably neither will the next one. But the one after that, maybe they will get to grow without seeing the current jobs situation as The One True Way. Which in the end means Schmidt got it right - we need universal access to education. That is literally the only problem that we need to solve - leave the future problems for the future people.

Strshps1MoreTim• 14 years ago

Insane comment ;)

There were quite a few stupid people saying things were impossible before. Don't worry, let your children worry is a sure recipe for failure. There are so many possible huge problems with AI, including an unfriendly one.

During human history the technology has played a huge role in determining the structure of societies. When the modern killing devices are cheap - bronze swords in Ancient Greece and Rome, muskets in per-revolutionary US and France - there is a lot of individual freedom, as the elite is incapable of ignoring the will of the masses. On the other hand when those devices are extremely expensive - knight armor in European Dark Ages - there is absolute control in the hands of a very few. Modern AI driven drones and robots are extremely anti-democratic, they allow a small elite to employ a few expert button pushers to destroy any kind of opposition. With CCTV, Facebook, WireTrap, Internet monitoring and logging, etc. your children will stand no chance to change the Orwellian future your stupidity has doomed them to.

martinced• 14 years ago

It's been 50 years or so that people are either waiting or fearing "AI".

Yet we're absolutely nowhere. Watson or Deep Blue may be "impressive" but they're not doing anything intelligent. No even remotely close to it.

"Siri: how can we solve the israelo-polestinian conflict while minimizing the number of human casualties?"

"Siri: how can I build a nuclear reactor that doesn't generate nasty byproducts and that doesn't go ape-monkey-shit should the sh!t hit the fan?"

I somehow really doubt I'll see real AI during my lifetime. OCR, finding patterns inside "big data", brute force searching, voice recognition, self-driving cars, etc. is not AI.

It's not because we can build "robots" mimicking stuff we do in an automated way ("I see a car coming at the right, let's slow down at this crossing") that we're getting closer to AI.

There's no thinking going on there.

I'm a bit surprised that a lot of people here seem to take it for granted that one day sentient Siri would answer like Leon when asked:

"Siri: what do you think about Asimov's three laws of robotics?"

; )

dagw• 14 years ago

"Siri: how can I build a nuclear reactor that doesn't generate nasty byproducts and that doesn't go ape-monkey-shit should the sh!t hit the fan?"

And when we get to a point where a computer can give an answer to that question we'll be dismissing it as "not real AI" since all its doing is pattern matching against a database of known experimental results and pre-programmed properties of elements and sub-atomic particles or whatever.

The reason we'll never see "real AI" is simply because as soon as computer scientists solve problem that used to be consider AI, it instantly becomes just another novel twist on a few obvious algorithms and the goalposts for Real AI will move a bit further down the field.

Eric Schmidt: How We Outrace the Robots

Related Stories