Zeynep Tufekci‘s recent piece “Failing the Third Machine Age: When Robots Come for Grandma” has been getting some attention. It’s a polemic against the prospect of using advanced technologies to provide elder care, embedded within a larger diatribe about technological progress, automation, and capitalism.

I don’t want to take on her big argument here. If you want a response to it, and a very different and mindfully optimistic view of tech progress and capitalism, check out our book The Second Machine Age (an age which, incidentally, Tufekci thinks we’re “close to concluding.” Erik and I think it’s just heating up). I want to concentrate instead on what she says about the idea of using robots to help take care of Grandma:

In my view, warehousing elderly… in rooms with machines that keep them busy, when large numbers of humans beings around the world are desperate for jobs that pay a living wage is worse than the Dickensian nightmares of mechanical industrialization, it’s worse than the cold, alienated workplaces depicted by Kafka.

… surely we should mourn if we put our elderly… in “care” of metal objects animated by software because we, the richest society globally the world has ever seen… cannot bring together the political will to remain human through taking care of each other, and making a decent living doing so.

Who could argue with such ringing language and humanist sentiment? I’ll start.

First of all, no serious person I know is advocating warehousing old people in rooms where they’re looked after only(?) by machines, any more than the Affordable Care Act set up death panels. Hyperbole is the ally of demagoguery, not of serious argument.

And to the extent that Tufekci is advancing a real argument, it’s one that stumbles both on its economics and its view of where the better angels of our nature spring from.

Her piece includes no numbers about the costs of elder care. And the numbers are daunting. Care.com reports, for example, that a semi-private nursing home room in Massachusetts costs more than $120k per year, and a 44 hour per week home health aide almost $60k.

Faced with these figures many families decide to take care of their oldest members themselves (a task that, as Tufekci notes, usually falls to women), but this approach has its own costs, and they’re also high. As reported by NPR in 2012, “a  MetLife report said that for the typical woman, the lost wages due to dropping out of the labor force because of adult caregiving responsibilities averages nearly $143,000.”

The direct costs do not go away if we decide to greatly increase public elder care; they just get paid by the government (and paid for via higher taxes or borrowing) instead of individuals. Maybe this is the right thing for us to do, but it wouldn’t change anything about how expensive it is to provide labor-intensive services in a high-wage country.

The one thing we can all agree on about automation is that it lowers the costs of good and services. Most people would also agree that cost reduction is what we need more than anything else right now throughout our healthcare system. Tufekci ignores these issues completely.

Instead, she paints a picture of how we will inevitably become desensitized to each other and therefore dehumanized as technology starts to help with caregiving tasks. I just don’t buy it, for the simple reason that technology has already taken over a great many things that used to take place between people, and humanity keeps chugging along admirably.

I don’t see the harm in creating technologies to help with elder care and more than I do with taking a pulse, connecting a telephone call, making a book recommendation, issuing an airline ticket, or doing any of the thousands of things that used to be done exclusively by people and are now done often by machines. I certainly don’t feel that because I now use technologies for these things my humanity is being chipped away at; I feel instead that having these things done better, faster, and cheaper by technology frees me up to have the kinds of human interactions that I actively choose to have. I’m happy to report that these have not gone to zero.

The idea that there’s some kind of clear Rubicon that we should never cross in the world of work — some tasks that should must always be done by people lest we lose something precious and essential about ourselves — is a perennially beguiling one, but it’s wrong.  Among its many flaws the deepest is that it doesn’t give us enough credit.

It suggests that we’re going to stop caring about our grandmas once we use technologies to learn if they’ve fallen, help them get out of bed, or yes, keep them company. Who here is not offended by that suggestion?


I’m on the brink of making a big real estate commitment in Cambridge, the idiosyncratic New England city I’ve called home since 1994. But a set of proposed regulations discussed at a License Commission meeting last night are so bad and so dumb they’re causing me to rethink whether or not I really want to live here.

The draft rules cover how people get rides in the city. And I’ll at least give the Commission credit for honesty; they’re not even really trying any more to cover up their assault on Uber, Lyft, and other on demand, smartphone-summoned ride services. The rules they came up with are quite clearly designed to drive these companies out of town; there’s really no other way to read them. As Uber summarized on its blog:

these regulations would:

    • Set a $50 minimum price for any non-taxi car ride, regardless of time or distance
    • Prohibit you from requesting a ride on-demand from anyone other than a taxi
    • Forbid any technological device from being part of fare calculation during a ride

Who else but a Cambridge taxi driver or regulator could possibly think this was a good idea?

I’ve been taking cabs here since I showed up as an undergraduate in 1984 (I wasn’t taking many back then because I was so poor). I’ve never had better than an OK experience, and I’ve had plenty of truly lousy ones: dirty, broken down cabs, unqualified drivers, and so on. I’ve also been unable to get a cab so many times I’ve lost count. When there are only 257 medallions for a city of more than 100,000 people this isn’t too surprising, but it’s always frustrating.

And then Uber came along, and life got better for me and all other ride seekers (I’ve never used Lyft, so can’t comment on it). I’ve always been able to find a car, it’s always showed up, and I’ve had an OK-to-great experience all but a handful of times. And each of those times Uber noticed my dissatisfaction (since users rate every ride), reached out to me, apologized, and tried to make it right via a refund. The Uber experience is so superior that I’ve basically taken ‘taxi’ off my mental list of ways to get around Cambridge.

The License Commission wants to put it back on not by improving taxis, but by taking away my preferred alternative. Can they be serious? This honestly seems like an elaborate piece of civic performance art in which I’m an unwilling participant, or an episode of “Nathan For You” where he somehow convinces bureaucrats to use their powers to maximum detrimental effect.

Uber supporters came out in force last night and the Commission backed off, saying that the public hearing “was the start of the conversation and committed to a transparent process going forward, involving both Uber and the people of Cambridge.” according to Uber’s blog.

My fellow Cantabrigians, let’s make sure we stay involved. There’s a lot riding on this issue.

aa-unemployed-college-grads-300x198Recent research continues to shed light on the big trends in the US labor market. Unfortunately, many if not most of them are bad news.

As Thomas Edsall describes well in his latest New York Times column, it looks like demand has slowed down for even the most cognitively demanding jobs (in other words, the highest skilled ones) since the turn of the century. During the 1980s and 90s demand for jobs had a U-shaped pattern: there were plenty of opportunities at the high end and the low end (thanks to the rise of low-skilled service jobs) and shrinking opportunities in the middle. My MIT colleague David Autor, who did the most work to document this phenomenon, calls this the ‘polarization‘ or ‘hollowing out’ of the US job market.

A new study by Paul Beaudry, David A. Green, and Ben Sand looking at the same phenomenon finds that a very different pattern holds since 2000, with demand for even the highest skilled jobs declining slightly. After looking at this research, Autor observed that the pattern has gone from a U-shape to “a bit more like a downward ramp” since the turn of the century.

This is bad news for several reasons. One of the most important is that the downward ramp appears to be leading to a “skills cascade” in which highly skilled / educated workers take jobs lower down the skill / wage ladder (since there’s not much demand at high levels), which in turn pushes less skilled workers even lower down the ladder, and so on. Larry Katz has found that “lots of new college graduates are moving into the service sector, that is, into traditionally non-college jobs, displacing young non-college workers.”

Where this all ends is anyone’s guess. Edsall quotes Larry Summers at the end of the article as saying that “The single most important step the U.S. government can take to reverse these discouraging trends is to mount a concerted, large-scale program directed at renewing our national infrastructure.” Erik Brynjolfsson and I certainly agree; infrastructure investment is one of main things in the ‘Econ 101 playbook’ for restoring growth and opportunity that we put forward in our book The Second Machine Age (along with other economic no-brainers like fostering entrepreneurship, reforming immigration, and investing more in basic research).

Edsall quotes our book in his article:  “the transformations brought about by digital technology will be profoundly beneficial ones. We’re heading into an era that won’t be just different; it will be better.” We do in fact say this (and still believe it), but it’s inaccurate to portray us as tech pollyannas or utopians.

The quote above refers to the uncontroversial fact that tech progress grows the overall economic pie (to use the economist’s favorite metaphor). However, we spend more than a third of The Second Machine Age highlighting that the distribution of this pie is becoming more uneven (and hence problematic) in large part because of this same progress, and discussing ways to mitigate or reverse this trend.

All the evidence I see these days solidifies the conviction that our labor market is shifting both deeply and quickly these days. How long before we take action in response? As Edsall writes, maybe policy changes will come more quickly once large numbers of college-educated parents see their college-educated kids unable to find the jobs that used to be regarded as essentially the birthright of their class.


600full-waterworld-screenshotI’ve been trying to figure out how to convey the scale of the ‘Big Data‘ phenomenon — the recent worldwide explosion of the volume of data encoded in digital form. Inspiration came from Randall Munroe’s fantastic “What if?” comics, which provide “Serious Scientific Answers to Absurd Hypothetical Questions.” (check out his 2o14 TED talk and pre-order the “What if?” book.)

So I decided to (poorly) imitate his methodology and try to seriously answer the question posed in the title of this post — if there was already an ocean of data in the world in 2007, how much more ‘datawater’ was there in 2013? I chose this metaphor of data-as-water because it’s a familiar one; we often use the imagery of having ‘oceans of data’ that we’re ‘drowning’ in.

I chose 2007 for no particularly good reason. It’s pretty recent, which makes any large increase in datawater more surprising. It wouldn’t surprise many people to learn that the amount of datawater has gone up a lot over the past thirty years, for example, but how much has it increased just over the past seven? 2007 is also when the phrase “big data” was just starting to be used.

Finally, it’s a year for which we have a couple estimates for how much worldwide digital data existed. The first is from the 2011 Science paper “The World’s Technological Capacity to Store, Communicate, and Compute Information” by Martin Hilbert and  Priscila López; 2007 is the last year considered in this research. The second is from a series of whitepapers published by EMC each year since 2007 that give the amounts of digitally encoded data worldwide.

These two estimates line up pretty well for 2007 (the one year they overlap), so I’ll use the Science estimate for 2007 and the EMC one for 2013 (the most recent year in the series). They tell us that there were 295 exabytes of digital data in 2007, and 4.4 zettabytes in 2013, giving an annual growth rate of 57.5% over the period.

I’ll associate the volume of digital data in the world in 2007 with the volume of the Atlantic Ocean, just because it’s the one closest to home. The Atlantic Ocean is big in absolute terms — it contains over 300,000,000 cubic kilometers of water — but it’s also relatively small; it makes up less than 0.03% of the volume of the Earth. So as I started my calculations I found that I didn’t have good intuition about how much datawater there would be in 2013. Would it be a thin film covering the Earth’s surface? Would it come up to our waists? Be deep enough to swim in?

It turns out that it would cover all of our creations and everything else on the planet to the point that we could happily boat around on the surface of the datawater ocean without worrying about bumping into any of the mountaintops of the former world. The volume of datawater created between 2007 and 2013 would cover the Earth to a depth of  84,417 meters (276,000 feet), which is almost ten times the height of Mt. Everest (you can double-check my calculations here).

The addition of this amount of water would certainly be the biggest change in the history of the world. Which makes it a pretty good analogy for the advent of Big Data.




Comparison of Nike’s successive sustainability reports reveals that the company used 106,000 fewer contract employees around the world in 2013 than 2012 (a greater than 9% drop), even as both profits and revenues increased by 16% and 5%, respectively. A story in the International Business Times states that this is because the company is “shift[ing] toward automation” even though it already makes most of its products in some of the lowest-wage countries in the world — shoes in Vietnam, Indonesia, and China; apparel in Brazil, Mexico, Argentina and India.

It could be that Nike’s overall contract employment went down because its product mix changed away from shoes and clothes and toward higher margin, lower labor offerings like the Fuel Band. However, this doesn’t appear to be the case. The company notes in its annual report “a shift in the mix of the Company’s revenues to lower margin geographies, products and businesses.” Furthermore, Nike has recently backed off from the Fuel Band and other wearables (and might soon drop them altogether). 

So it appears that instead of chasing ever-lower labor costs around the world, the world’s largest sportswear maker is instead turning to automation. Nike’s financial results show that this strategy is paying off so far. Anyone think that what this company is doing is the exception, rather than the rule and the shape of things to come?

If so, challenges are looming for developing countries.

Dani Rodrik and others have highlighted the phenomenon of ‘deindustrialization‘ around the world — the growing tendency of developing countries to not go through a period of heavy employment in manufacturing. In the past, countries like Korea, Taiwan, and China raised up many of their people by having them make stuff (i.e. by industrialization).

If Nike’s example really is part of a larger trend, however, this path to prosperity might become less available in the future. Companies will use automation rather than cheap labor to make their goods, in which case the workers who would have supplied that cheap labor are left out. As Rodrik writes,

Less room for industrialization will almost certainly mean fewer growth miracles in the future… Given premature deindustrialization, today’s developing countries will have to travel different, as yet unknown, and possibly bumpier paths to democracy and good governance.

Erik and I keep saying that offshoring is a way station on the way to automation. Nike’s recent experience with contract labor looks like a data point in support of that contention. I predict that there will be others.

The research is clear that technological progress has greatly benefitted people in the developing world so far. I wonder, though, if automation and deindustrialization might be creating a ‘silicon ceiling’ on growth — a situation in which even low wages are no longer an attractive alternative to technology. If so, the global shift away from labor and toward capital will only accelerate.

Every so often I have an experience as a consumer so bad that I have to write about it. The latest were the interactions I had with Comcast to get a new TiVo box working properly at my Mom’s house in Chicago and then (having, I thought, learned from that one) ordering a Comcast DVR from the company itself along with a bundle of services and getting someone to come to my house to install it.

Screenshot 2014-04-24 16.46.55

The only way to get through these experiences was to treat them as tests of my patience, tell myself that they were great blog fodder, and try to see the humor in them (as if I were the King of Swamp Castle instructing his guards; this honestly helped). I encountered impenetrable web sites; dropped calls; having to explain what I wanted over and over, sometimes to the same person; inconsistent information  (I had a person in one chat window telling me I wasn’t eligible for a particular package at the same time someone else in another chat window was signing me up for it); and a truly noxious combination of apathy and confusion.

Comcast is certainly busy these days. It’s trying to merge with Time Warner Cable to create a giant quasi-monopoly that will, according to a letter opposing the deal written by Netflix CEO Reed Hastings and CFO David Wells, “pass over 60 percent of U.S. broadband households…with most of those homes having Comcast as the only option for truly high-speed broadband.”

Hastings and Wells are concerned that “The combined company would possess even more anti-competitive leverage to charge arbitrary interconnection tolls for access to their customers.” That sounds right, but I have another worry: that the company just won’t care very much about taking care of its customers, because where else are they going to go?

We’ve seen this before. The late 19th and early 20th centuries were a time of large industrial enterprises, some of them run by ‘robber barons’ who, as Robert Heilbroner points out in his essential book The Worldly Philosophers, “were uninterested in producing goods.” (emphasis in original).

Railroads, which like cable companies can wind up looking very much like local monopolies for a lot of reasons, provided some great examples for Heilbroner. As Jay Gould and Cornelius Vanderbilt were fighting for control of the Erie Railroad, a superintendent wrote that “The iron rails have broken and laminated and worn out beyond all precedent until there is scarcely a mile of your road, except that laid with steel rails, between Jersey City and Salamanca or Buffalo, where it is safe to run a train at the ordinary passenger or train speed.” At the time the golden spike was driven to connect the Northern Pacific Line, “the capitalization [was] far ahead of what it should be for what there [was] to show, and the selection of routes and grades [was] abominable. Practically it would have to be built over.”

So the financial machinations continued and occupied a great deal of time and energy from the empire builders at the top of the company, while the actual business crumbled. Sounds familiar, and helps me understand how Comcast can win Consumerist’s Worst Company In America contest twice in the past five years. Maybe I should stop complaining and consider myself lucky that the cable entering my home actually delivers signal. But I can’t help wanting more, and wondering how I can get it. Any ideas?

Next Monday morning at Noon I’m talking with king of all media Walter Isaacson about The Second Machine Age at the Aspen Institute in Washington DC.

Screenshot 2014-04-16 14.33.31

I give Walter that title because in addition to editing Time and being CEO of CNN, he’s also written runaway bestsellers (about Steve Jobs) and Pulitzer Prize nominees (about Henry Kissinger). In his current job as the head of the Aspen Institute he gets to exercise another of his great skills: conducing interviews in front of an audience.

I’ve had the chance to listen to Walter talk on stage with many people over the years at the Aspen Ideas Festival, and always walked away feeling like I’d been challenged and learned something. Now it’s my turn to face his gentlemanly grilling.

Space at the event itself is limited but our conversation will be livestreamed; click here to watch it on Monday. It’ll also be made available online afterward.

I’m really looking forward to this one. If you turn in, I very confident you won’t be disappointed or bored.

Just a quick note to let Washingtonians know about two happenings in their city this week. I find them interesting not because I’m speaking at them, but rather because of the people I’m speaking with.

Tonight from 5:30 to 7 I’m sitting down to talk with Prof. Amitai Etzioni at George Washington University about “the effects of the coming digital revolution;” the event is free and open to the public, and details are here. I’m going to use this opportunity to play interviewer myself and ask Prof. Etzioni how a communitarian thinks about the second machine age and the opportunities and challenges it brings. If you’re geekiness extends to the subjects of technology, policy, economics, and/or political philosophy, I think you’ll enjoy the conversation a lot, especially because it’ll include audience members.

Screenshot 2014-04-10 11.19.55

On Saturday afternoon the IMF, as part of its anniversary celebration, is bringing together an all-star panel to talk about “IMF 70 Years Later: Reflections and Looking Ahead.” The Fund’s managing director Christine Lagarde will participate, as will Madeleine Albright, Peter Ho from Singapore, and my friends Bob Gordon and John Lipsky. Ali Veishi will moderate (I wonder if he’ll try to stoke the ongoing debate about tech progress and innovation I’ve been having with Bob or suppress it).

We’re at George Washington University’s Lisner Auditorium from 3:45 to 5 on Saturday, and a live webcast will also be available. If you’re nearby please do come; I can promise you an informative, informed, and lively discussion.

As these events show, the conversation in our nation’s capital is starting to include the topic of rapid technological progress and its consequences. I find this very good news…

MIT in NYC, April 4 2014

Just a quick note to let New Yorkers know that the MIT Initiative on the Digital Economy is coming to your city this Friday, April 4.

MIT Sloan faculty, alums, and other digital movers and shakers will come together from Noon to 7 pm on Friday at the TimesCenter (I haven’t been inside the facility yet, but it looks gorgeous).

We’re almost at capacity but I’m told there are a few seats still available. If you’re interested, register here by April 1 (no fooling).

If you’re at all interested in technological progress and its implications for business, society, and economics I think you should join us. Rather than hyping the event, let me just post the agenda; I think it more than speaks for itself.

Hope to see you there!


12:00 p.m. Registration begins
The Gallery, Main Level

12:30 p.m. Welcome Remarks
The Hall, Lower Level
David Schmittlein
John C Head III Dean, MIT Sloan School of Management

12:45 p.m. The Second Machine Age
The Hall, Lower Level
Erik Brynjolfsson, PhD ’91
Director, Initiative on the Digital Economy, Schussel Family Professor of Management Science, MIT Sloan School of Management

1:30 p.m. Coffee Break and Move Rooms
The Hall & Gallery, Main Level   

2:00 p.m. How Business Models Will Have to Change
Andrew McAfee, ’88, ’89, LGO ’90
Co-Director, Initiative on the Digital Economy, MIT Sloan School of Management

2:25 p.m. Panel Discussion: The Disruption of Everything
Moderator: Andrew McAfee, ’88, ’89, LGO ’90, Co-Director, Initiative on the Digital Economy, MIT Sloan School of Management
Panelists: George Colony, Chairman & CEO, Forrester Research, Inc.
William H. Janeway, Managing Director, Technology, Media, and Telecommunications, Warburg Pincus
Hilary Mason, Data Scientist in Residence, Accel Partners

3:10 p.m. The Science of Social Media Influence 
The Stage, Main Level  
Sinan Aral, PhD ’07, David Austin Professor of Management, Associate Professor of Information Technology and Marketing, MIT Sloan School of Management

3:35 p.m. Coffee Break  
The Hall & Gallery, Main Level 

4:05 p.m. The Social Physics of Cultural Innovation: Keeping up with the Machines
Alex ‘Sandy’ Pentland, PhD ’82, Toshiba Professor of Media Arts and Sciences, MIT Sloan School of Management; Director, Human Dynamics Laboratory; Director, Entrepreneurship Program, MIT Media Lab

4:30 p.m. Fireside Chat with John A. Thain, SB ’77, Chairman and Chief Executive Officer, CIT
Moderator: Erik Brynjolfsson, PhD ’91, Director, Initiative on the Digital Economy; Schussel Family Professor of Management Science, MIT Sloan School of Management

5:00 p.m. Panel Discussion: Automation and the New Mind-Machine Boundary
Moderator: Erik Brynjolfsson, PhD ’91, Director, Initiative on the Digital Economy;Schussel Family Professor of Management Science, MIT Sloan School of Management
Panelists: Zoë Baird, President, The Markle Foundation
Carl Bass, President & CEO, Autodesk, Inc.
Thad Starner, SB ’91, PhD ’99,

5:45 p.m. MIT’s Initiative on the Digital Economy
Erik Brynjolfsson, PhD ’91, Director, Initiative on the Digital Economy; Schussel Family Professor of Management Science, MIT Sloan School of Management
Andrew McAfee, ’88, ’89, LGO ’90, Co-Director, Initiative on the Digital Economy, MIT Sloan School of Management

6:00 p.m. Closing Remarks 
David Schmittlein, John C Head III Dean, MIT Sloan School of Management

6:05 p.m. Cocktail Reception
The Gallery, Main Level

Embedded image permalinkJust a quick note to let New Yorkers know that I’ll be in Manhattan tomorrow (Wednesday, March 26) participating in a very worthwhile event: a discussion and reception sponsored by tech industry advocacy group FWD.us on technology, the middle class, and the American dream.

It takes place at AOL headquarters at 770 Broadway. I’m not sure what time doors open, but there’s a reception before our panel starts at 6:30. FWD.us founder Joe Green moderates (I’ll vouch that he’ll keep things lively), and I’ll be joined by Van Jones and Scott Murphy.

FWD.us initially took up immigration reform, a tremendously important topic, and has now broadened its brief to include the issues I’ve been studying and writing about: the interplay among technological progress, aggregate economic growth, and individual economic opportunity. I’ll talk about our book The Second Machine Age as well as what we’re learning from the most recent research (including Thomas Piketty’s great new book Capital in the 21st Century).

I hope you’ll come, ask a question during the discussion, and stick around for the reception afterward. I don’t think you’ll be sorry you did…