Erik Brynjolfsson on the Second Machine Age
Feb 3 2014

Erik Brynjolfsson of MIT and co-author of The Second Machine Age talks with EconTalk host Russ Roberts about the ideas in the book, co-authored with Andrew McAfee. He argues we are entering a new age of economic activity dominated by smart machines and computers. Neither dystopian or utopian, Brynjolfsson sees this new age as one of possibility and challenge. He is optimistic that with the right choices and policy responses, the future will have much to celebrate.

RELATED EPISODE
Andrew McAfee, Megan McArdle, and Lee Ohanian on the Future of Work
Andrew McAfee, Megan McArdle, and Lee Ohanian talk with EconTalk host Russ Roberts on the future of work. Recorded before a live audience at the 33rd Santa Barbara Economic Summit, the conversation begins with each participant making a brief set...
EXPLORE MORE
Related EPISODE
Amy Webb on Artificial Intelligence, Humanity, and the Big Nine
Futurist and author Amy Webb talks about her book, The Big Nine, with EconTalk host Russ Roberts. Webb observes that artificial intelligence is currently evolving in a handful of companies in the United States and China. She worries that innovation...
EXPLORE MORE
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.

READER COMMENTS

Tom
Feb 3 2014 at 10:27am

The account of our culture given by the guest speaker in this week’s ET was shockingly thin – nothing more than industrial production. Ford, Gates, Jobs, Bezos. These billionaires are the constellation to guide thought.

This is the mindset of flyover country. Culture, in this world view, is industrial production, there to be continually transformed, in faster and faster cycles, by new heroes of production and selling.

This is a recipe for mass destruction of all values.

rhhardin
Feb 3 2014 at 11:07am

The explosion of machines is due not to artificial intelligence advances but low hanging fruit with the advent to really cheap computers.

Anything that’s simple to automate gets automated because machine cost is no longer an obstacle.

With artificial intelligence, when a new problem is solved, it turns out not to haev been artificial intelligence but the discovery of a mundane mindless algorithm that covers the cases.

The fantasy of AI though has been strong for a long time. I think it’s a male idea.

Brian
Feb 3 2014 at 5:46pm

This was a great podcast. I took one exception. The part where professor Brynjolfsson stated that:

“By some measures the generation graduating today is the first one in American history that is not more educated than their parents are.”

I find that impossible to believe. My nephew is taking, in public school, computer classes that weren’t available when I was a kid. The math that is expected is much more advanced. A friend of mine, who is very proficient in math, recently went back to school and had to take a 1st year pre-requisit and was blown away by what kids out of highschool knew.

Steve Sedio
Feb 3 2014 at 8:44pm

As mentioned, the idea of a self driving car was realized well before we expected to. Computer power, combined with the web, has allowed brute forcing machine “evolution”. Google provided a free 411 service to develop / refine voice recognition. I can see a similar process in vision systems based on facial recognition.

Therefore, I predict it will also take less time than expected for technology to solve, and surpass human agility and dexterity. Janitors and barbers beware….

Computers can handle chess, which IMHO requires a combination of creativity and analysis. Computers don’t have creativity, but are able to throw enough spaghetti at the wall that something “creative” sticks. Add in that computers already shift through journals looking at key words to identify new relationships for evaluation. Who says people have to do that evaluation?

Even without a singularity, computer power increases far faster than we can grow well educated 20 year olds.

The other thing to realize is, the end of work as we know it doesn’t have to occur in the next 10, or even 100 years. Shy of some global catastrophe, the tech evolution will continue.

I can see a time in the future where little work will “need” people. As I’ve said before, the problem will be the transition.

Assuming tech continues to reduce the cost of living, my suggestion is we minimize unemployment by periodically shortening the work week.

lloydfour
Feb 3 2014 at 8:46pm

Edgar Rice Burroughs wrote a scifi novel where the cars on a highway were fast and efficient to get you cross town in minutes. The highway congestion disappeared. Another problem arose, it took two hours sitting on the entrance ramp waiting for a space to open on the highway so you could accelerate onto the highway.

Jessease
Feb 3 2014 at 9:27pm

How we come to think about technology and automation in the long-term will depend on how society deals with an ever-shrinking labor pool resulting from those same technological advances.

I don’t believe the government or free market economy will be able to continue in their current form under the growing cycle of poverty that results from job destruction. As we start entering an era of Peak Jobs in which quality labor prospects begin to diminish for most people, society will need to rethink wealth creation and monetary circulation in a growing people-free workforce.

If we continue on our current path of debt and denial, I see it leading to mass unemployment or underemployment and the social strife that arises under those conditions. If however we alter that path and get it right, it would likely mean shorter work weeks with living wages and more free time. After all, shouldn’t that be the societal reward for successfully developing these technologies?

lloydfour
Feb 3 2014 at 10:19pm

I really liked this podcast.

More scifi–imagine a society that has huge increases in wealth from massive use of robots so that the society can choose to pay the price of increased inefficiency by limiting the use of robots and increasing the use of humans. Ironic that robots become the source of their own replacement by humans.

cswaters
Feb 4 2014 at 6:58am

“By some measures the generation graduating today is the first one in American history that is not more educated than their parents are.”

Above Brian wrote he didn’t believe it. I have the same opinion. This is flat out wrong no matter how you look at it.

The NyTimes did an article and put up an interactive map in the middle of the last year (easy to find with google) showing a sharp increase, across the board, in the number of americans with college degrees. Stories like that come out monthly. I’m 34 now and for the last twenty years I’ve heard this sort of flap doodle in one form or another. It’s all riffs on the same theme, “This generation will be worse off than their parents… civilization has hit its peak and it’s all down hill from now on.”
I wasn’t around in the 60’s or 70’s but I’d bet the same sort of nonsense was being spewed then as well.

Craig
Feb 4 2014 at 7:45am

Russ: You have a great chart in the book where you show that in 1972, median and average income start to diverge. Before that, they are very similar. Then all of a sudden they diverge dramatically. The average keeps growing, and the median starts to just totally slump. Do you find it plausible that something happened in 1972 that caused this incredible discontinuity in the data? It’s rather shocking, really. … But the median person just kind of chugs along. In fact, 1972 is just this big discontinuity. You can’t really–it’s hard to explain. I’m not saying I have the explanation, by the way.

One thing that happened in 1971 was Nixon closing the gold window, the elimination of the last tie between the dollar and gold, that permitted and resulted in a discontinuous change in the inflation of the money supply. I suggest that since then, in effect, the government has been, by this means, confiscating and redirecting the entire sum of economic progress that would have previously gone to raising the standard of living. The result of this is stagnation for most people and growing relative wealth for the early recipients of the newly created money (the Cantillon effect).

Mr. Econotarian
Feb 4 2014 at 5:07pm

Russ, I know you must have decided that to skip correcting Brynjolfsson on the “dismal science” reference.

Of course it was coined by Thomas Carlyle in his 1849 work “Occasional Discourse on the Negro Question” where he argued that slavery is morally superior to relying on the market forces of supply and demand, and he labeled the profession of economists who disagreed with him, including John Stuart Mill, as the “dismal science,” since Carlyle believed that the emancipation of slaves would leave them worse off.

Russ Roberts
Feb 4 2014 at 5:21pm
David Zetland
Feb 4 2014 at 9:37pm

First of all, I LoLd on this: “in terms of the higher paid jobs, whether it’s creative writing or software or art”

Where ARE these high paid creative writing and art jobs? I have about 2,000 friends who’d love to partake.

Second, and more serious, I agree with other commentators that the returns to the elite and destruction of “mind” jobs will leave very little for the masses. Although some may “end it” by not having kids or suicide-by-booze (as we saw in the ex-USSR), I’m guessing that most will be around. For them, it makes sense to tax the rich (I favor property taxes) and subsidize the “working poor,” but I’d go further (a la Keynes) and give people a basic income subsidy (as recently discussed in CH), as that would “free” them to pursue paid or unpaid work that was fulfilling — and probably not threatened by robots.

GregS
Feb 4 2014 at 10:31pm

Great podcast. I have a borderline irrelevant comment:

“In terms of what I think is practical for most people, with the possible exception of Ph.D. economists: calculus is just widely needed. But that’s sort of a tangent.”

Was this meant as a pun? If so, bravo. If not, it made me chuckle anyway.

bogwood
Feb 5 2014 at 6:48am

What are the energy flows required for the new technologies? The author owes us at least an estimate. Is it like the Tesla,an ecological disaster? The Tesla fails even with boundaries set very narrowly,the embedded energy in the battery. Where does the self-driving car fail? The education of the mechanics children? Or does it have a net return?
The economy is a subset of the environment and by assuming free environmental services it needs to be a small subset. Economists are a subset of ecologists and by analogy are possibly undertrained by two orders of magnitude. The author provides no ecological data.

David McGrogan
Feb 5 2014 at 8:19am

I enjoyed the podcast but as with the Tyler Cowan podcast at times I found myself wondering, “And people at the low end are just taking all this lying down?”

I’m no Marxist but it’s easy to imagine greater moves towards redistributive taxation, unionisation, and so forth, with populist politicians at the forefront, as a backlash on the part of the lower classes against a technological revolution that leaves them behind.

I’m not a sociologist either, but big technological change often correlates with big social change. The era of the industrial revolution also happened to be the era of revolution, too. I’m not predicting anything, but it is worth pondering.

Ron Crossland
Feb 5 2014 at 8:28am

Terrific podcast. While there are variations in opinion concerning what is AI and measuring education levels over time, I thought the general discussion was very important. (A tangent on the education idea is the popular discussion today that perhaps a college education isn’t worth the economic investment for many.)

Would love a future podcast that talked more specifically about how politics and economics (to a lesser degree) is overwhelmed by the rules and thinking of the past. We are in a second machine age and yet we talk about and make political and economic decisions as if we are still in the early to middle portion of the first machine age.

Picking up the theme of different/better economic measures that inform how the world is changing rather than how the world has been would extend and drill down on ideas from this podcast.

D. F. Linton
Feb 6 2014 at 11:58am

I was surprised that the discussion seems to focus exclusively on the absolute advantage of computers over people in some area rather than on comparative advantage.

JasonW
Feb 6 2014 at 7:58pm

agree that entrepreneurship is more important than ever. When that topic came up and it was asked what discourages it, I immediately thought access to capital.

The great thing is that technology has provided a solve for that little thorn in the entrepreneurs side which is crowd sourcing. Websites like Kickstarter have unlocked a whole new world for people who have an idea but not the money to make that idea happen. If you’re looking for an example, look at the Pebble which launched the smartwatch revolution.

my second thought in response to people’s resistance to change is to provide an economic, functional argument for Christianity which teaches us to live as if our life and our choices are not completely our own and to work on each problem that comes our way as if it were a unique puzzle presented by God to teach us how to live the life he wants us to live. By living that way, you become immune to change. Easier said than done, I know.

Anonymous
Feb 6 2014 at 9:58pm

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.–Econlib Ed.]

Daniel Barkalow
Feb 7 2014 at 1:04pm

I think the step we will have to make at some point is to give people jobs doing things that don’t create value but are difficult and satisfying for them. We’ll be in a situation where need isn’t an issue, because there’s enough stuff produced that, if people turn up with the money to buy it, they still won’t create shortages. Then (in Voltaire’s words) you still need to do something about boredom and vice. But there’s no reason someone’s job couldn’t be to learn Icelandic (for example). Learning a language keeps you out of trouble and makes you feel good about your accomplishments, and it can’t be done by a computer (or a human who isn’t you). Sure, it’s not productive, but that’s no longer a big deal because of the productivity surplus. The hard thing, of course, is getting to an economic system where the increase in value from machine turning raw materials into useful objects somehow gets distributed to people who are diligent about trying to learn languages.

(As I side note, I think a great part-time job for many Americans would be to get an appropriate amount of exercise on a regular schedule. Even if it didn’t produce anything, it would reduce health care costs, as well as feeling like work to the people involved.)

Arnim Sauerbier
Feb 8 2014 at 11:45am

One overlooked factor absorbing those put out-of-work by productivity increases/machine-substituion is the increasing number of people employed by government.

Regulation, taxation, police, domestic and global spying, military hardware production, overseas military presence (in approx 135 countries) are all growth “industries”.

Do these expenditures just absorb the surplus or do they have net economic benefits? Do you see anything on the horizon that will reverse this trend?

lloydfour
Feb 10 2014 at 6:47am

Saw on news show that a researcher has succeeded in attaching an artificial arm to the patient’s brain. That is the future, not robots. Humans will be enhanced and will out perform robots. Borg anyone?

Reason Able
Feb 15 2014 at 10:58pm

Our tax system is very skewed, in a sense punishing people for creating jobs, because that’s what we tax. And one of the first rules of economics is if you want to discourage something, you know, you raise the price; you put a tax on it.

I am surprised that as intelligent as Dr.Brynjofssons he makes this argument that if you tax people’s income you will discourage people from working. I expect this to come from republicans/libertarians who know nothing more than supply & demand not Dr. Brynjolffson. The fundamental assumption that everyone works for money and nothing else so if we put tax on the income they will work and earn less is in the face of it flawed. There are people who work or start business to make money and no other reason but I suspect this is a small portion of it. Most people start a business because they are passionate about what they do. And, money is only a side factor. You need some minimum amount for you to live and survive but beyond that it is because you enjoy what you do.
If his statement is true then no one in corporate world would want a raise because that means they will have to pay more in taxes. So far, I have never seen anyone decline a pay increase because their labor is going to be taxed at a higher marginal rate.
One place this could happen is if the marginal tax rate is 100%. Even when marginal tax rate was 90% people wanted to make more money only that they figured out how to avoid paying taxes using loopholes.
It is best if all economists lay off of this argument because it really demeans all workers.

John
Feb 18 2014 at 3:33am

A couple of thoughts:

I do not think that it is a modern day phenomenon that someone who is really talented can reach a greater audience. Take, for example, Joseph Haydn, the music composer of the 18th Century. His music was played by most European orchestras and captured a 70% “market share” in some English Venues during his lifetime. So I don’t think that the internet is creating something new here, other then perhaps amplifying it in some degree for a small number of artisans but not for all people with great talents – like plumbers, engineers who can only apply their knowledge locally.

Also, I think that economist should know better to say that productivity is increased because a teacher can reach a greater audience. That measure is not productivity. For a teacher to increase his/her productivity he/she would have to teach a greater amount of material in a shorter amount of time with the same of better student understanding and retention of the material.

As far as smart cars: the reason that they would be cool is the idea that they would be networked together and thus enhance the flow of traffic during rush hour traffic as rush hour traffic exists today. However, the future machine vision has only a few people working with machines doing the rest thus making such a smart car obsolete… in that no one would be commuting to work anymore.

My greatest problem with all this is that no one can really know what the future is going to be. The best solution is to create a system that has the greatest amount of flexibility that can adapt the quickest to what ever happens. Such a system does not come about top down and we are now a long way away from having the freedom necessary for the coming changes.

John
Feb 19 2014 at 1:11am

I am surprised that economist consider wikipedia ‘free’ and contrast it with the encyclopedia sold door-to-door in the past. If one delineates the cost of purchasing a computer and the monthly ISP costs, wikipedia is anything but ‘free’.

Consider purchasing an encyclopedia set in 1950. A salesman came to your door and said you could purchase the entire set for, lets say, $500. Amortized over 20 years that comes to ~$2.08 a month.

Now jump to 2014. If you purchase a MAC you will spend $1000+ and the ISP cost per month is roughly $50. That means over 20 years (4 computers) the amortized cost per month is ~$66.66. Now I realized that not all of that cost is looking at wikipedia but $2.08 < $66.66. Perhaps if you multiplied the $66.66 into the exact percentage that you do use wikipedia you might find that the costs are the same and therefore no gain has been made at all in terms of cost reduction. Thus how can economist make such blanket statements about wikipedia and consider it 'free'? Even if you use the computer at the library the cost isn't free when you consider the taxes that go to pay for the library.

Comments are closed.


DELVE DEEPER

About this week's guest:

About ideas and people mentioned in this podcast episode:Books:

Articles:

      • "Economic Possibilities for Our Grandchildren," by John Maynard Keynes. 1930. Reprinted in Essays in Persuasion, New York, W. W. Norton & Co. 1963, pp. 358-373.
      • Milton Friedman. Biography. Concise Encyclopedia of Economics.

Podcast Episodes, Videos, and Blog Entries:


AUDIO TRANSCRIPT

 

Time
Podcast Episode Highlights
0:33Intro. [Recording date: January 16, 2014.] Russ: Now, you argue that technology has really taken off in the last few years. What are some examples that you think have either happened already that are significant or are about to happen, and why are they important? Guest: Frankly, it's caught me off guard at times, because just a few years ago I was teaching my students about things that machines could do well and things that they couldn't do well. And one of my favorite examples was driving a car through traffic. And I said, there's no, like, program or algorithm for that; that's something with too much sensory perception and uncertainty. And then lo and behold about a year ago I was riding down Route 101 in a driverless car. So I was proven wrong. And I try to be a little bit of an optimist about technology. But it's not just Google's car. It's also IBM's Watson that plays Jeopardy and beats humans. And if you know anything about the game 'Jeopardy' there's all sorts of unstructured questions and puns; there's geography, there's math, there's poetry; there's popular culture. It's quite a smorgasbord of information. For the first time in history we can talk to our machines, our phones, and they understand what we are saying and they'll carry out our instructions. That's quite a milestone. There are amazing things happening with big data. And many of these breakthroughs really were in the past 5 or 10 years or so. Russ: Are they going to make a difference? I have Siri on my phone. I find myself using it from time to time as a novelty item, but it's not part of my life. We don't have a relationship. The Academy Award nominations just came out, and Her is getting a lot of acclaim. But right now it's kind of a novelty item. The driverless car is really cool. Is the driverless car, to take one example--is it going to happen or is it just going to be an experiment? Guest: After riding in it, I'm pretty convinced that it's more than just an experiment. And I gave you examples that are deliberately sort of on the leading edge. But all of them, I think, are going to become more and more mainstream. To take the case of the driverless car, I don't think we are going to be seeing in the next decade cars driving through Boston traffic--people can barely do that. But I do think we'll see cars driving down the highway in California. When I was riding down there, I felt, frankly, safer than I did with a human driver. I went through sort of three phases. When I first sat in the car, I was a little scared, you know, because the car in front of us on the highway came to a dead stop. And I was hoping that Google's driverless car would recognize that and stop. And indeed, it came to this nice smooth slowdown and stop. And then it accelerated again as the other car picked up. And then the second phase I went through, I was just kind of exhilarated; and Andrew McAfee and I--the co-author of the book--we were literally waving at other cars we drove by, and just very excited. But then I got to this third phase, within about 10, 15 minutes, where I was actually kind of bored. Because, you know, we drove all the way from Mountain View up to San Francisco and back again; and the car--it kind of drives like your grandmother might, very slowly; always drives 55, doesn't break the law. And it's very smooth. And you know, after a while, you are sort of like: Okay, I get it. It's kind of like watching a dishwasher work. It's just not that exciting any more. And so I could easily see that technology widely used for sort of like super-cruise-control. Or maybe you'd pull it up to a parking lot, you'd drive through the city but when you got there, you'd get out and say, 'Okay, car, go park yourself.' And a few specific applications like that will become common within a decade. Russ: I'm sorry you were bored on the ride. You could have been listening to EconTalk, or other forms of-- Guest: That's right. I could have been reading. Russ: And I've had the good fortune of driving in Boston over the winter break. It's a very unpleasant experience. Why do you say that that is unlikely to happen? Because I would think that's the most important economically and life-changing way that a driverless car would work. The idea of taxis that you could use as a service, like Uber, to summon a driverless car to take you to the airport or to take you across town would be--that would be extraordinary. You don't think that's coming? Guest: Well, it's going to depend on the conditions. And this is where I rely on my smarter and more knowledgeable colleagues at the MIT (Massachusetts Institute of Technology) Computer Science, the iLab and elsewhere, who are actually building driverless cars and autonomous robots. And one of the joys of my job is that I spend time interacting with them, visiting their labs and asking them, 'What's doable? What's not? Is this smoke and mirrors or is what I rode in for real?' And they tell me that things like the highway driving, certainly very doable. Driving in sort of wide, easy, low-traffic conditions, very doable. But as it starts getting more and more complicated--driving through Manhattan or Boston during busy times--you have to use, still, a great deal of judgment. I think actually my impression from driving in Boston is that if you really did follow all the laws and rules, you wouldn't move, because there are just too many exceptions[?]. You wouldn't be able to get out into traffic without kind of bending the rules a little bit. And it would be hard with current technology or even the next decade's technology, to make those levels of judgments. I suspect that will come as well. But with the driverless car like so many of these other things, it's not all-or-nothing. And sometimes the first 90 or 99% of the problem is very doable, but that last 1% will take longer. Russ: A friend of mine once tried to get to the Museum of Science in Boston, which is particularly difficult. A lot of one-way streets, strange, non-rectangular grid there. He finally gave up and he asked a policeman how to get there, and the policeman gave him the directions: 'Take that second left,...'. And my friend said: 'Isn't that second left going the wrong way down a one-way street?' And the policemen said, 'Take a chance.' That's not going to be a driverless car's strategy, as you point out. But the thing I do wonder about, is what the world would be like for a driverless car in a world of other driverless cars. Because the driverless car today is designed to drive with other drivers. If they were all driverless, or most were driverless, there might be some really even better things possible, like moving at high speeds in tandem, etc. Anybody working on that? Guest: Oh, absolutely. You are exactly on target. It actually becomes much easier when you allow other driverless cars to be the ones that you interact with. Sebastian Thrun, who, as you may know, headed the Google driverless car project, was telling me his idea was instead of this high-speed rail that they are thinking of putting between San Francisco and Los Angeles for a fraction of the cost. For a few billion dollars you could dedicate a lane on the highway for driverless cars. And it would be so easy for him to provide the technology that you could do that. Now, it wouldn't have exactly all the efficiencies in some ways, but on balance it would be more cost effective. And as you suggest, it could have the cars basically train themselves--become trained. They link up. They could drive within a few inches or feet of each other. Most of the pavement on a congested highway right now, when you think of a highway as being totally full, actually about 90% of the pavement is unused because the lanes are twice as wide as a car, and it's typically about 4 or 5 lengths between cars. Even in heavy traffic. So you could get an enormous increase in capacity if you allow the cars to know about each other and assume that the other cars are also driverless, at least on some of the lanes. Russ: It's a good argument for selling the highways, because a private entrepreneur would love to implement that. The government, less so. We'll see. Could happen. It's a fabulous idea.
9:23Russ: You started off by saying that you used to tell your students there are some things you are good at and some thing machines are good at. We do see this creeping capability of machines--which sounds somewhat ominous; not necessarily, but it could be-- Guest: It should be good news if we do it right. But certainly the way we have things set up now, it can also be bad news for a lot of people. Russ: So, where do you think that line might stop? What's going to be left for us? Tyler Cowen, recently a guest on the program, suggested we'd be coaching--would be one of the things we could do, because machines wouldn't do that very well. And we'd be working with machines, maybe, helping them. But that's a pretty narrow slice, and he's somewhat pessimistic about the future. What do you think will be left for us as artificial intelligence advances? Guest: Well, first off, as your last sentence suggested, it's a constantly advancing line. So 3 years from now, 5 years from now, 15 years from now the line is constantly moving and often faster than I expect. Sometimes slower. So in the shorter time frame there are still plenty of things that people can do. The ones that Tyler mentioned actually are a good place to start. Motivating other people and interacting with us--we've been hard-wired over hundreds of thousands of years of evolution to know how to interact with other people. And there are a lot of subtleties we can pick up on that have been very hard for machines to be programmed to pick up on. So, we're still pretty good at that, and for that matter, fine motor control is another thing that we have a big evolutionary advantage. Picking up a penny or a dime is something that most 3-year-olds can do, but robots have an incredibly hard time doing that. Fine motor control to do those kinds of tasks, or even the gross motor control to walk around a room. If you've seen some of the Atlas robots--we can talk about that later--that are in the latest DARPA (Defense Advanced Research Projects Agency) grant challenge, they walk very, very slowly and clumsily compared to a person. So there's a set of jobs from gardener to barber to janitor that are actually relatively immune in the short term. There are also emotional connections and there are a lot of creative tasks as well. So, at the often higher end, in terms of the higher paid jobs, whether it's creative writing or software or art or perhaps parts of economics, I suppose, that involve a lot of intellectual feats that we still haven't figured out how to embed in a machine. So, those are some of the categories that right now, I think there's still growth; and there will be for some time. Russ: So one of the obvious categories you mention is software--programming machines, developing new machines. But how does that--let's look way down the road. Not too far--not a thousand years--but maybe 30. Let's go 30, 40 years. We've made some pretty serious transitions. What do you imagine life is like for the bulk of, say, the American society? In other words, is it going to be a world where a handful of people--again, this is how Cowen has suggested--who make very large amounts of money designing and being the creative force behind this world that we are talking about, and the rest of us are cutting each other's hair? It sounds pretty depressing. Nothing wrong with being a barber. But most of us don't aspire to it. We have other loves and interests. So, what's out there? What do you think it will be like? Guest: Actually I think that's a very plausible scenario. We talked about--Andy and I talked about that a little in our last book, Race Against the Machine. We talked about how people should try to race with the machine. People like data scientists get paid a lot in part because they are complementary to cheap computation, cheap data; and that makes their skills very valuable. People who write great software are very complementary with machines, and they can see their talents leveraged a thousand-fold or a million-fold, and that can make them very, very wealthy. And that's great as well. But in a superstar economy, you often get these winner-take-all markets. You don't necessarily want every Tom, Dick, and Joe's software for tax preparation. There's one or a few that tend to dominate the market in most categories, and that means that it is hard for everybody to be the world's best in some category. Or at least any meaningfully large category. I think there will be growing inequality unless we take some actions and maybe there will be more difficulty for people with average skills to find jobs. As we said in the last book, and we also have a chapter devoted to it in this book, there is really probably no better time to be a talented entrepreneur, but there is also no worse time to be a person with no special skills or aptitudes.
14:43Russ: So let me try to create a positive vision for that depressing part. Let's say that I'm one of the 90% or 95% who can't program those machines, can't come up with the newest robot, can't make the robot better at picking up nickels, pennies, and dimes. So, I'm a gardener or I'm a trash collector--there's a whole range of things that may be either humans can do them better or they are cheap enough that it's worthwhile to do it. If everything is so inexpensive because of this productivity, maybe we enter the world that you allude to in the book, and we've talked about this on the program before, Keynes's essay on his grandchildren--we're in a world where maybe I work a few hours a week. But my command of goods and services for that few hours is enormous. And those few hours are not important. So instead of spending 25%, 50% of my waking time at work, I'm spending a fraction of that and I have a richer life. Is that a possibility? Guest: It's definitely a possibility and it's one that we should probably try to work towards. As you suggest, Keynes's prescient essay, "Economic Possibilities for Our Grandchildren," he described exactly that kind of a world. And he was very much right on target when it came to extrapolating growth rates; it turns out he was good at doing exponential math. So he got our living standards pretty close to where they actually are, and in part we are fortunate that we have continued to have that kind of exponential growth, since 1930. But where he was really quite wrong was how we spent our time. I think in his era he imagined anybody as wealthy as you and I are today, we would just want to spend all our time as English lords did, maybe sipping tea and occasionally listening to poetry and hunting foxes, I guess. But people continue to work quite a bit. I think that's in part because we've discovered so many new goods and services that there are more interesting things to do than sip tea in a kind of chilly castle. And so, you know, we are motivated to do other things. Also I suspect there is a sociological component. I'm not a sociologist, but people seem to like doing productive things. I was recently talking to Bob Putnam. He's a researcher at Harvard who described how devastating it can be to communities when work disappears, when they lose their jobs. And it's not just a matter of income. It's a matter of family structure, of crime, teen pregnancy, drugs, a whole set of ills that come along with that. To home in more specifically on your question, I am hopeful that we can have that kind of an abundant economy; and we do talk about, in one of the chapters, all of the free goods and services that are available and the many cheaper goods and services. But it's also possible that the wages for many kinds of workers will fall even faster than the price of goods and services will fall. As you get this growing dispersion. And for those people it may not be such a pleasant outcome. They may not be working not because they are satiated but because they are close substitutes for machines and machines can do what they do at a lower and lower price. And I think ultimately there is always a price that clears the market, but for some people that price may be $0.50 an hour or less, and that's not a very pleasant existence. Russ: But as you point out in the book, a lot of what we enjoy these days, we enjoy without paying for it. EconTalk I like to think is one of those things. Guest: Is it free? Good. Russ: We don't charge our listeners. Anybody can listen. We support it as an educational cause. So there are a lot of things in life like that. So in that case your purchasing power--later we are going to talk about what this does to GDP (Gross Domestic Product) because you talk about it in the book--but you just think about this image of 'oh, my goodness, my wage is falling faster than prices are falling'--when prices hit zero, that's a pretty good deal. And so if I get satisfaction from browsing Wikipedia and contributing to Wikipedia, say, instead of my job, it might be an okay life. I don't know--I'm trying to be optimistic here. Because it's kind of scary. Guest: Yeah. Let me build on the optimistic part and then give some caveats. We do have a chapter on free goods and we talk about Wikipedia. It turns out I've done some research on the value of free goods in the economy, and the consumer surplus is probably on the order of $300 billion a year from just the Internet free goods. Russ: Explain what consumer surplus is. Guest: Sure. Well, consumer surplus is the difference between people value a good--what they'd be willing to pay if they absolutely had to--and what they actually do pay. So, in the case of, say, Wikipedia, we get that for free, whereas Encyclopedia Britannica costs thousands of dollars. By many metrics, Wikipedia is better--more comprehensive, approximately the same error rate. So the people who used to pay for that encyclopedia are now getting something for free, and that difference is an increase in their consumer surplus. And you can do that for every good. Almost any good that you buy, almost by definition you are getting some consumer surplus, because, since you bought it, you must value it at more or equal to what you paid for it. The total consumer surplus in the economy is enormous. And it, as you suggested, is quite different from the GDP of the economy. The GDP is what the actual market prices of the goods and services multiplied by the number of goods and services, the amount of goods and services. So those are two different kinds of numbers. And economists tend to focus, I think, much too much on GDP. Those are the numbers that are reported out to 7, 8, 9 significant digits every quarter; and the Press trumpets them. But think about it. If the price of something goes down, like that encyclopedia, that tends to lower the GDP even though it's not making us worse off. It's making us better off. So we should probably care more about consumer surplus than we do about GDP. Russ: It's just a lot harder to measure so it tends to get pushed aside. Guest: That's exactly it. It's a case of focusing on the things we can measure even if they are not the things that matter to us. One of the things we talk about in the book is how we need to reinvent our metrics and our measures. But just to make sure I close the loop on that point about the free goods--they are tremendously valuable and we calculate that in my research and we discuss it in the book. But I also want to point out that not everything is getting cheaper. Not everything is a digital good that is subject to those kinds of economics. So, education, health care, housing--these are some pretty big ticket items that, depending on how you quality-adjust them, are still big, big chunks of our spending and aren't subject to that same kind of economics of 'free'--at least not yet.
22:24Russ: You say 'at least not yet'. It seems to me that if we got out of the way a little bit, we could unleash some of that. The education system being at least one example of the potential for giving people very low cost access compared to what we currently use as the technology of a person standing in front of 25 people in a room. We could really make some big changes. Guest: We could make a huge difference there. I think that you are right on target with education. If you look at the other industries--media and retailing and manufacturing and finance, they have all been revolutionized by this digital revolution. But my industry and your industry, it's barely been touched. We stand in front of a classroom much the way Socrates did thousands of years ago, perhaps even with a similar piece of chalk and slate. And MOOCs [Massive Open Online Course] and other digital technologies have the potential for having two big breakthroughs. One you alluded to about really scaling things. Someone like Sebastian Thrun who I mentioned earlier, when he taught his artificial intelligence course at Stanford, he moved that onto a digital MOOC and reached hundreds of thousands of people. And that was a huge increase in his productivity and a huge increase in quality for a lot of people who then had the opportunity to learn from him. But the second benefit is subtler but I think is going to be even more important, and that is when you digitize things you tend to dramatically improve their measurability. And we've just been talking about the difficulties of measuring things. But education hadn't been measured very well, not just in terms of the test scores but in terms of the real micro level, what's working and what's working. We have a MOOC at MIT that my colleague Anant Agarwal taught, Circuits. And he was able to measure how people were listening to the lectures and doing the problem sets on a minute-by-minute basis and keep track of things. And he was surprised to find, for instance, that most of the students in the class started working on the problem sets way before they listened to his lecture on the particular topic that they were supposed to be working on. Russ: Sometimes there's too much information. Guest: Yeah, there was too much information. And what they'd do is they'd work on them, and then when they got stuck, then they'd go to the lecture and maybe listen to the part that they needed to listen to, to help them with that part of it. And he realized that that's the way people really get motivated--when they have a problem, then they go seek the knowledge. And it flipped around some of his assumptions. That was just one of the insights that he got, but he got many, many insights about that, about what learning styles work and don't work, and even how they differ from one kind of student to another kind of student. And that's ultimately not only going to raise the level of our educational quality, but I think it will put it on a different trajectory. It will increase the slope so we can improve it year to year or even week to week compared to what we do now.
25:20Russ: Let's talk about the current state of the economy for the average worker, average person's standard of living. You are worried, as many are--I am less so--about two things: stagnation and inequality. They are often conflated, which I think is a terrible intellectual mistake. You suggest, as do many, that the standard of living for most people hasn't budged for the last 20 or 30 years potentially, because at the top it's getting larger; and they are capturing, allegedly the gains. Inequality is also getting worse. So, make the case for that and suggest what it might tell us about what we have going forward. Guest: Sure. I mean, in many ways this was part of the catalyst for Andy and I to write this book. Tyler Cowen--is he a colleague of yours? Russ: He used to be. I used to be at George Mason U. Now I'm at the Hoover Institution. Guest: He wrote a very inspiring book called The Great Stagnation, pointing out this issue about stagnating median income. And it's not just him but there were many economists such as at the American Economic Association meetings in Philadelphia and had what amounts to sort of a debate with Bob Gordon of Northwestern U. who wrote an article-- Russ: Also a stagnationist. Guest: Yeah. His article was on the growth of debt or the end of innovation--something like that. And Dale Jorgenson of Harvard, who had some pessimistic views. And I was sort of the lone optimist on it. But there are a lot of economists--I guess there is a reason they call economics the dismal science--who have a pretty dim view of what's been happening recently to median income. Obviously employment has suffered a great deal over the past decade; it doesn't seem to be improving. If you look at the employment-to-population ratio, as you may know, the recent increases, improvements in the unemployment rate, mostly reflect people dropping out of the labor force, not so much new people getting jobs. So those are some dismal and discouraging statistics. But at the same time, Andy and I were confused because we spent a lot of time with these technologists at the MIT Media Lab and CSAIL--the Computer Science and Artificial Intelligence Lab. Silicon Valley. You never meet such optimistic people as these guys. There's no problem they don't think they can solve. And they tell us that innovation has never been faster and technology will solve all our problems, even going as far as the Singularity and all. So, these two different views, these groups really didn't talk to each other, and we want to try to reconcile them. And the way we reconcile them, and we talked about it in the book, is that yes, technology is advancing incredibly fast. We are in a very innovative period. But there's no economic law that says everybody is going to benefit evenly from improvements in technology. Or even that everyone is going to benefit at all. It's possible for some people, even a majority of people, to be made worse off, not just in relative terms but in absolute terms. Now that's not something that's happened in the past, or for 200 years that didn't happen. But in the past decade or two I think there is growing evidence that that has been happening. We have been seeing the pie growing quite large. We are at record wealth--$77 trillion dollars is the new record we hit last week. Record productivity. Record GDP. Record profits. And it is I think growing very fast, not all of it measured. But the official data suggests that median income is in fact pretty stagnant and probably lower than what it was in the 1990s. And 'median' is not the same as average, of course. Median is the 50th percentile and those below. So, those are some statistics that illustrate how technology can be growing the pie but not necessarily helping out everybody. Russ: Yeah. I don't believe the statistics, at least not totally. I'm willing to accept the possibility they are right. But when people talk about them, they never accept the possibility that they could be wrong, those statistics. So, just to take an example, and it's one that you really cover in the book: in 1979 or 1989--people use different dates because they cherry-pick different government measures of income. But typically they are looking at, say, household income. They ignore the fact that household structures changed dramatically since 1979, say. There has been an enormous increase in the divorce rate, and enormous delay in marriage, a lot more people are single; households then have fewer workers. And as a result, when you compare median in 1979 to 2009, say, it's not something close to an apples-to-apples comparison. And so, as you point out--is there anybody in 2009 who would like to be back at their 1979 income? Have their 1979 income and prices compared to today's income and prices? Their standard of living is higher for an enormous portion of the folks. You can debate how people feel about inequality, but it's obvious, as you point out, that quality is poorly measured in the price data. And this household demographic effect distorts the measurement. Guest: I think you are asking exactly the right questions and going after some of the serious flaws with those metrics. On the household measure, that is a flaw. And that's why it's important to correct for household size, or it can be useful to correct for household size. That probably eliminates about 20-30% of the discrepancy, if you do that. And there are other things you can do, in terms of adjusting for benefits, of the quality of health care, and potential biases in the Consumer Price Index, which I personally believe is biased and our standard of living is growing faster than it says, although people like Bob Gordon will argue the other side of that-- Russ: Fair enough-- Guest: And he'll point to things that people used to get on the other side that we don't get today. And you could debate--because it is a little hard. You'd like to do an apples-to-apples comparison but it's almost physically or philosophically impossible, is a better word--metaphysically impossible to do that comparison. Because they are different people, different goods, different services. So can you say whether having your gas station attendant wipe your windshield is more valuable than having access to Siri on your iPhone? I don't know. There are just some weird comparisons that you can never really do. But what I think there is little dispute about is that what has measured the gap between the top 1% and the 50th percent or whichever percentile you want to focus on, has grown. And if the median income hasn't fallen, it hasn't grown a whole lot. It may be flat. So I think that, you know, depending on which tweaks you want to make, you may move it around by a few percent maybe one way or the other. For instance, I tried to correct it for the free goods on the Internet; and there's tens of billions of dollars of additional value every year. But when you add them in it's not enough to come close to bringing the rate of growth of median income up to what it was in the earlier eras. So, I mean, I think there are some people, like you, who are skeptical of the numbers; and I think there is justified skepticism. My sense is that the majority of economists would agree that median income, however you measure it, probably hasn't grown much since the 1990s. There's no theoretical reason that it had to. So I tend to be with the majority on that. But I think it's something that requires continuing further study.
33:41Russ: Let me ask a different piece to this puzzle. You have a great chart in the book where you show that in 1972, median and average income start to diverge. Before that, they are very similar. Then all of a sudden they diverge dramatically. The average keeps growing, and the median starts to just totally slump. Guest: Right. Russ: Do you find it plausible that something happened in 1972 that caused this incredible discontinuity in the data? It's rather shocking, really. Right? There is one explanation, which is, the standard explanation is: Well, the average person, that includes the high outliers--as you point out in the book, it's not a normal distribution--so the high outliers pull the average. But the median person just kind of chugs along. In fact, 1972 is just this big discontinuity. You can't really--it's hard to explain. I'm not saying I have the explanation, by the way. Okay, go ahead. Guest: I don't think it's that hard to explain. First off, it's not like there's a one-time shock in 1972 that didn't continue. What we saw was they started diverging then and they've continued to diverge ever since then. And if you look at a lot of different metrics--we have another chart where we look at education and how the salaries of people with different levels of education. And before about 1970 or so people with high school education or college education or post-graduate education all saw their wages grow at about the same rate. And afterwards they started diverging. And by the 1980s, people with high school or less education were falling in absolute terms, whether they were male or female. And there are a lot of other ways of slicing the data. And they all tell a kind of a similar story, that things are diverging. And in the book we have three complementary explanations for what's going on. The first is that widely studied by economists called 'skill-biased technical change,' which is this idea that--and you alluded to it earlier I think--that technology tends to be complementary or favor those with more education or more skills by various measures metrics and tend to substitute or lower the demand for people with lesser skills. And that has accelerated and helped drive some of that. Especially actually in the 1980s and 1990s. The second explanation is 'capital-biased technical change'--the change, the technology can also help one factor of production versus another, like capital versus labor, and the share of income in the U.S. economy that has gone to capital has gone up, and the share that has gone to labor, including compensation and including for that matter CEOs (Chief Executive Officers) and superstars. But if you add them all together, that share has gone down. Not because the CEO's salaries have gone down but because of the compensation going to the other 99% of Americans has gone down more rapidly. And the third change is what we call 'superstar-biased technical change', and that is really honing [sic] in on that top not-even 1% but one hundredth of 1%, you know, the superstar baseball player, Scott Cook who wrote TurboTax and is a billionaire and the musicians and many others who have come up with skills or talents or luck that allow them to have a product or service that is not scalable to millions of other people. And that's been very, very rewarding to them and has created a tremendous amount of value for the consumers of that product; but it also means that people who have the second best or the tenth best service in that category aren't in this much demand. You put those three things together--skill-biased technical change, capital-biased technical change, and superstar-biased technical change--and you can have a pretty big divergence in the incomes of different kinds of people.
37:51Russ: Yeah, and I--you know, I certainly agree that all of those kinds of things are important. Just a side note on the educational fact, and that will let us segue into some policy issues. So, it's true that if you only go to high school, you don't go to college--worse, if you don't finish high school--that the prospects for those folks, the economic/financial prospects for those folks have really deteriorated, especially relative for people who go on to college or go on to graduate school. Guest: Right. Russ: The problem with that analysis is it's not a random sample. The proportion of people who go to high school, who only finish high school, or more dramatically the proportion who don't finish high school, changes over time. And so the people in the college group, or the people in the don't-finish-college group or don't-go-to-college group--it's a very different group of people. We treat those educational categories as if they are all the same. They are not. The human capital level and the intangible, unmeasured skills of somebody with, say, a 10th grade education today versus a 10th grade education 50 years ago--those are very different people. Guest: Well, let me make two points on that. First, in concept I think you are exactly right. That's certainly something that is well understood by Larry Katz and David Autor and others who do it. And their models include by changes in supply and demand. And they do their best to account for the changing composition of the people and how the demand effects as well as the supply effects change. That's a conceptual point. The other perhaps more distressing point is that while you are right that that composition was changing quite a bit for most of the 20th century, in the past couple of decades, really, we have kind of hit a ceiling, or we've stopped improving that ratio. By some measures the generation graduating today is the first one in American history that is not more educated than their parents are. So it's not like we are continuing to see a big change in that. America's greatest idea, we write in the book, was mass education. And that was a response to the first machine age--was widespread education of people to win this race against technology. The race between technology and education. But in the past couple of decades, we are not winning that race any more. The United States whether it is by international measures or measured against ourselves in the past have not been having anywhere near the investments and benefits in education that we once had. Russ: Fair enough. And to make your point stronger, if you increase the proportion going to college, and college is less educational, you get a separate problem. Same is true with high school. I think the average high school graduate of today is not as skilled in the--the unobservable parts of that person that can't be measured are different, but so are the inputs that were actually applied and given to them in the educational experience. In some dimensions. In other dimensions--I don't know. Math is pretty serious, still in American high schools. It's hard to dumb down calculus. It can be done! But a lot of kids still do take serious math. And as you point out, we struggle sometimes internationally, but we do still learn something in school. Guest: My pet little thing, I just wanted to mention, is I'm not as much of a fan of calculus as I once was, and I'm on a little push in my high school to replace calculus with statistics. In terms of what I think is practical for most people, with the possible exception of Ph.D. economists: calculus is just widely needed. But that's sort of a tangent. Russ: Well, it's interesting. My wife is a math teacher, and she is teaching a class of seniors this year, split between calculus and statistics, for one of the levels of the school. And statistics is--I agree with you. Statistics is in many ways much more useful for most students than calculus. The problem is, to teach it well is extraordinarily difficult. It's very easy to teach a horrible statistics class where you spin back the definitions of mean and median. But you become dangerous because you think you know something about data when in fact it's kind of subtle. Guest: Yeah. But you read newspapers saying--I just grimace because the journalists don't understand basic statistics, and I don't think the readers do either. And that's something that appears almost daily in our lives. I'd love it if we upped our education in that area. As data and data science becomes more important, it's going to be more important to do that.
42:36Russ: So, let's take the pessimistic view, which is that what we are observing now is a transformation of the Second Machine Age where technology increases the size of the pie but the share going to the bulk of the people isn't very large. What can we do down the road to make that better? What are some of the policy recommendations that you think are important? Guest: Well, fortunately, we've already talked about one of them, the very first one that we focused on, which is education. And I don't mean just investing more in education, although that wouldn't hurt, but I mean mostly reinventing education and bringing digital technologies to it. Let's not dwell on that too much more because we've already spent some time on it and I know you have a limited time for this podcast. The second big suggestion and set of suggestions we have is around entrepreneurship. We need to create a more dynamic economy. And our point about entrepreneurship which is sometimes misunderstood is not that we think everyone is going to be or should be an entrepreneur. It's not like that's what you do when you can't find a regular job. We want to push for more entrepreneurship and make it easier in different ways, because entrepreneurs are the people who are in charge in our economy, are inventing new products and services. They are the ones who are inventing new jobs. I mean, it's not an MIT professor who is going to be able to figure out what the hot new products are or a government bureaucrat or even the President. It's something we need to crowd-source. And the way we crowd-source that is we have a thousand or 10,000 or a million entrepreneurs try different ideas. A lot of them are going to be really dumb and they are going to fail. Some of them are going to be breakthroughs. You know, when the First Machine Age automated so much of the economy--we used to have 90% of Americans who were in agricultural; now it's less than 2%--all those people didn't just become unemployed. What happened was that, you know, ingenious people like Henry Ford, Steve Jobs, Bill Gates and others helped invent entirely new industries, new products and services that employed people. And yes, there were some preconditions. They had an educated population to work with and other things that they could draw on. But we have not been doing very well in that recently. You may be surprised that the entrepreneurship and the growth of young new companies is less in the past decade than it was in the 1990s or 1980s. And when we lose that dynamism we lose the ability to replace the jobs that are being automated away. Russ: What do you think is stopping that? Why do you think that has changed? Some of it is obviously cyclical, related to the Great Recession. But you suggest in the book there are some other factors. Guest: Yeah. Even before the Great Recession, it happened. You know, I've talked to the people who have studied this most closely, at the Kauffman Foundation and people like John Haltiwanger, and they are actually quite puzzled about what's going on. There are some regulations, often at the state and local level, that protect incumbents--whether it's the taxi drivers that are trying to keep Uber from flourishing in Cambridge and other places, or-- Russ: France[?]-- Guest: Or whether it's nail salons. Yeah. So that's part of it. There may be that people are more fearful, and if there is not a safety net they are afraid to try new things right now. Until recently health care was not very portable. And I know people who wanted to start jobs but they had to stay with a big company--sorry, new business but they had to stay with a big company to preserve their health insurance. We put a big tax on labor right now. Our tax system is very skewed, in a sense punishing people for creating jobs, because that's what we tax. And one of the first rules of economics is if you want to discourage something, you know, you raise the price; you put a tax on it. So these are all possible reasons. But it's something that certainly requires a lot more attention. Russ: We do have a great culture, which has been a great advantage for America in the past. That's not going away for a while, so it should continue to help us. Guest: Well, there's another one: part of that culture is an immigrant culture. And in Silicon Valley more than half the companies are started by immigrants, and we've become much, much welcoming to immigrants than we used to. Just earlier today, I was working on trying to deal with the visa and green card issues of a recent Ph.D. Actually, he's holding off getting his Ph.D. because once he gets it the clock starts ticking and the government kicks him out. I mean, what kind of a policy is that, when you get your brightest, most educated people in the world to come here and then we kick them out? So we should probably staple a green card to every Ph.D. diploma and encourage them to stay and welcome them not just to Silicon Valley and Boston but all over the country. So immigration policy and the culture that has that openness is something we could improve as well. Russ: There's a lot of fear. There's a lot of fear about the future, and I think we've touched on it indirectly, right? If you think that immigrants come in--I'm excited for that Ph.D. to come; I'm also excited for the uneducated person to come from Guatemala, open a restaurant, do whatever that person can do. Right now people who immigrate to the United States often work very hard, often work multiple jobs. But a lot of people are worried they are going to drive down the wages. They are like machines, just really cheap machines. Guest: Let me just take up on what you said. The fear--I think you are exactly right. This fear unfortunately often pushes us to do counterproductive strategies, like discouraging immigration when we should be encouraging it. And more generally trying to preserve the past at the expense of the future, when really that's just not going to be possible. We can't stop technology, we shouldn't stop technology. What we need to do is embrace the dynamism that helps us adapt to that. The more we do to try to slow down change, I think the more stagnant we become and the worse off we become. The only successful strategy is one that embraces change and does what it can to benefit a broad set of people. Russ: One of the optimistic ways to look at the future is to say, Well, I don't know what jobs are going to replace the ones that are going to be made obsolete by the new technology, but that was true in 1950, in 1900, in 1850. And they came along, as you said, bright creative people found new things for people to do. I think the worry--those of us who are optimistic, we have to be honest about it--is, and you point this out in the book: it's one thing when John Henry gets replaced by a machine. So, John Henry had muscle. And a machine is better at muscle than a human. But when it's brains, you start to wonder what is left. I think the [?] things, I don't know what they are, so that's not very comforting to most people. It requires a little bit of a leap of faith. Guest: You and I are very much in the same space. I think the analogy is that the First Machine Age replaced muscle; and then we went on to doing mostly more cognitive tasks. In fact, control systems, cognitive tasks become more valuable. They are complements to increased power and ability to move the world. There is no point in having a steam engine or internal combustion engine or any power system if you don't have somebody directing what it should do. So that increases the value of human labor. But when you start automating cognitive work, then it gets a little harder. It's not so obvious what the human complements are. Maybe they are more substitutes. We talked a little bit about relationships and those kind of skills and maybe there are some kinds of areas where people have an E.Q.--an Emotional Quotient--that's better than the machines. But even that--I have friends and the Media Lab like Cynthia Brazeal--I don't remember whether she made it into the book as we talked about it over the summer. They are working on these emotive computers that pick up on subtleties. Right now they are working with autistic kids. They have far more patience than most humans do working with them. And some of them are listening in on call centers, like Sandy Pentland, also at the Media Lab, has one that listens to the tone of voice of callers at call centers and warns when they are about to explode perhaps because the person in the call center realizes it. So it's not clear that we'll also have an advantage in that area, either.
51:21Russ: What about longer run solutions? What do you think is possible? You talk about some of the ideas in the book. Where should we be thinking, in not just the next few years? Guest: The best thing is to get people working. In the book we talk a little bit about the basic income and negative income tax that Milton Friedman and Richard Nixon actually were big advocates of, which is a surprise to a lot of people, as well as a lot of liberals. But we actually aren't as eager to hand people money, not because it doesn't solve some of their economic needs--it does. But as Voltaire said, work addresses three great ills: boredom, vice, and need. And it's not enough to just address the need. You have to address the others. We talked about Bob Putnam earlier and what happens when work leaves the community. We'd like to do what we can to keep work in the community. And so in our long-term recommendation chapter we talk about a growing investment in something like the Earned Income Tax Credit, which basically subsidizes work. Instead of the minimum wage, where the payment is made essentially by the employer, in essence penalizes somebody, the Earned Income Tax Credit is something that everybody contributes to; and even if the worker doesn't have a marginal product enough to make it worth paying them $10 or $15 an hour, you could still bring their income up to that level and keep them working through an Earned Income Tax Credit. And that combined with more education, more entrepreneurship, better matching of people--something we also talk about in there--I think for a good period of time, 10, 20 years or more, can keep a lot of people working, maybe as many as are now, in all sorts of productive endeavors. And if you go to the real long, long future then maybe we get something like a Star Trek economy, where a lot of our basic needs are taken care of, where, in the words of Keynes, we 'solve our economic problem'--not in terms of all of our wants and wishes but in terms of our basic needs, like food, clothing, and shelter. And that would be an economy where robots, machines, take care of a lot of that drudgery. Andy and I sometimes call that a digital [?]--the Athenians during their golden era spent a lot of time discussing philosophy and writing plays and going to plays and poetry. And they were able to do that because they had slaves. They were human slaves doing all the drudgery work. Maybe at some point even in our lifetimes we will have more and more robot slaves doing that drudgery so that we can devote our time to leisure and listening to podcasts like yours. Russ: I actually think we are a long way toward that. Another part of my optimism is to suggest that we've mismeasured the quality of life right now, not just where we are headed. An enormous portion of people--it's not 70%, but it's not 5%--a large portion of people have an immense amount of leisure on the job. I recently asked people to contact me if they listen to EconTalk at work, and I got well over 100 emails--I forget how many I got--but people who confessed or bragged or just let me know: Well, I'm doing this sort of mindless stuff; I listen to EconTalk. And we do so much--not everybody, for many, many people the quality of life-- Guest: That's right. Those are Amazon's prime shopping hours as well, during working hours. People are clearly--the jobs are nothing like the painful drudgery that they had a century ago, or even 50 years ago. I think on balance people's work lives are much more flexible--not just in terms of the leisure, but even when they are working, I think the nature of the work tends to be more interesting, in part because the deadening routine parts of it are the easiest ones to automate. Russ: They are gone. And, although it's awkward to wear a toga at work and be that Athenian in the 21st century, when you are working from home, as many people do now, you can wear a towel, even. You don't even have to go up to the toga. Guest: I won't ask you what you are wearing right now. Russ: I'm fully dressed. I am working from home. But I'm fully dressed. Last question: are you optimistic about the human enterprise? Not about technology per se, not about the American economy in 12 years, but about the human enterprise, as this technology world that we live in just does explode, and it's going to? Guest: I'm optimistic but with a huge qualifier. I started off telling you about these two groups that I interact with that don't talk to each other--the dismal economists and their pessimism--not all of them, not you, but many of them--and the over-exuberant Utopian technologists, who think that technology will just solve all our problems. And I think they both make the same mistake. They assume that technology is destiny and they plot out, this is going to happen, get used to it. Although I loved Tyler's book and I think he's got a lot of insights, that would be one place I quibble, is that I think he's a little too resigned to a particular future. Andy and I call ourselves 'mindful optimists,' because we think if we take action and we adjust our skills, our organizations, even our economic institutions and the way we organize work, we can adapt to technology. And the technology will race ahead. And we don't need to try to slow it down. What we need to do is speed up our adaptation to it, just as we did in the First Machine Age. In that case it was widespread primary education; it was the invention of new kinds of organizations and new ways of working. I don't know for sure what all the adaptations are going to be moving forward. We have three chapters in the book where we have recommendations where we lay out some of our ideas, but we also try to be humble and say that people are going to discover other ideas and encourage others to email us with their suggestions or post them on our website, their suggestions. My experience has been that generally people can make those kind of adaptations, and the biggest obstacle to that is getting the diagnosis right. And with this book we are trying to get the diagnosis right about what the issues are. And if we do that I think we'll be able to get the prescriptions right. Unfortunately a lot of people in Washington and in the public discourse I don't think are grappling with this really important issue. I think it's the grand challenge that our society is going to face going forward, but if we do grapple with it, at the end of the day I'm a mindful optimist. Russ: Washington seems to be very focused on the next 6 months. Not necessarily the most helpful attitude. Guest: That's right. They've got these sort of short-term [?]. But at the end of the day, I don't know much about Washington and politics, but one friend there told me, 'You know, Erik, the thing about Washington--they are not leaders; they are followers.' That's kind of what they are designed to be. And if the American people demand these kind of changes, then they'll follow; they'll want to get out in front of that crowd. But it's a mistake to think they are going to be leading things in the way, say, technologists in Silicon Valley do.