How Safe Is Your Job? A New Freakonomics Radio Podcast

Listen now:

PC Creative Destruction

Economists preach the gospel of “creative destruction,” whereby new industries — and jobs — replace the old ones. But in this era of technological wonder, has creative destruction become too destructive?

That’s the question we ask in our latest podcast, “How Safe Is Your Job?” (You can subscribe to the podcast at iTunes or elsewhere, get the RSS feed, or listen via the media player above. You can also read the transcript, which includes credits for the music you’ll hear in the episode.)

It’s an expansive conversation with economists, historians, a roboticist-turned-cartoonist, and a special guest who is not quite human. You will hear from:

+ David Autor, an MIT labor economist who has written well on employment growth and technological change. While acknowledging that a lot of middle-class jobs are vanishing, and that most job growth is among low-wage workers, Autor is generally an optimist: “[T]he interactions by which technological changes lead to changes in employment are really rich and complex,” he tells us, “and it’s not simply a matter of, you know, a machine does the job, therefore the worker doesn’t do the job, therefore there are fewer workers needed.”

+ John Komlos, a retired professor of economics from the University of Munich who has written a working paper called  Has Creative Destruction Become More Destructive?”. His answer to that question is a resounding “yes.” Unlike Autor, Komlos believes that the current technological revolution, unlike past revolutions, is a net job destroyer — and worse. “Innovation is not going to give us nirvana or a just society or a good economy,” he says. “The ‘selfie moment’ destroyed Kodak, which at its peak employed 145,000 people. And these were mostly middle-class jobs, you see. Apple is employing 47,000 people and that is an iconic company today. And it’s not going to be a job creator in the next foreseeable future. Facebook is employing 7,000 people. That’s it!”

+ Randall Munroe, creator of the webcomic xkcd and author of What If?: Serious Scientific Answers to Absurd Hypothetical QuestionsMunroe, who used to work in a NASA robotics lab, entertains our question about the likelihood of a robot apocalypse. (Good news: it’s unlikely.)

(courtesy Randall Munroe)

(courtesy Randall Munroe)

You’ll also hear the story of the rise and fall of the piano industry, a story that is relevant in many ways to our current economic patterns. It’s told by the historian Richard Lieberman and the N.Y. Times reporter James Barron, each of whom has written a book about the Steinway piano company. Lieberman’s is called Steinway and Sons; Barron’s is Piano: The Making of a Steinway Concert Grand.

Not that anyone cares, but David Autor assures me that my job — writer, podcaster, etc. — is perfectly safe. But a pesky interloper named Jasmine (who has visited our podcast before) disagrees.

Anthony N

It may be due to editing or the summary nature of the discussion, but the podcast/David Autor seems to suggest that there is something inherent in the nature of a particular "sector" where increases in productivity is either directly or inversely related to employment, and he uses agriculture versus medicine or law.

I'm not sure that it's something inherent to the industry; rather, it's the nature--scalability--of the kind of productivity improvement. Currently medicine and law are as labor intensive as agriculture, since seeing any one patient or drafting a legal brief requires time and effort that's only incrementally improved by word processing or electronic medical records; one still can only service one client/patient at a time for any given hour of work.

However, I don't see that universally being the case, and you can look at areas like radiology in medicine or legal research, and they are ripe for the kind of scaling that suddenly allows a handful of developers to replace thousands of man hours of traditional legal or medical work.



A note on creative destruction: I am an eternal optimist and love to embrace new technology, but since I moved to Rochester, NY I can no longer deny the terrible downside of our digital revolution. Once home to Eastman Kodak and thousands of blue collar jobs, Rochester is now place where the income disparity is overwhelming. We have one of the lowest graduation rates/highest child poverty rates in the country. No downtown tech hub is going to address the lack of blue collar work available.


I wonder if Kodak was really destroyed by what Dr. Komlos calls “the selfie moment”, or by Kodak's refusal to keep up with new technology. After all, companies like Canon & Nikon seem to be doing quite well. Wikipedia lists about 40 active digital camera makers: And digital technology expands the 'camera' concept to cell phones, webcams, rear-view cameras for your car... It'd be interesting to compare employment in all these to peak film camera employment.

Nor, for that matter, has the horse industry gone away. The domestic horse population has (per Google) gone from about 25 million in 1925 to 10 million today, but that's still a significant industry. Add to that the fact that nowadays horse owners are more likely to buy horse trailers and trucks to pull them.


I would like to point out to Autor that sex does, in fact, "require learning new skills and mastering them." Not everyone finds sex "enjoyable immediately and forever."

Ben C

Stephen's followup question at 28:20 demonstrates his skill as an interviewer.

On matters of public importance, economists have a great deal to offer by virtue of rigorous statistical analyses. But frequently, assertions are untethered from any empirical basis and economists smuggle in (deliberately or not) assertions that carry a philosophical valence. The danger is that when economists do this, it's easy to credit the philosophical assertion with an undeserved cloak of authority or expertise.

Professor Autor is a great example. He has real insights that are based on his research. But he gets intellectually sloppy when he starts making grand statements about the human condition (i.e. what gives life meaning). Without Stephen's followup, I would've been tempted to nod along instead of recognizing that Professor Autor's thoughts on the matter had no more support than the thoughts of a random person on the street.



Here's a question I rarely see asked and never seen answered: As the demand for automation increases, the supply side will see increasing competition. The only real way to differentiate as an automation engineer is to do your job faster and cheaper (which in their case generally means creating superior tools and techniques). As this continues, we may eventually reach a point where automating any given job is both faster and cheaper than job training. If this happens it does NOT MATTER if we run out of ideas or not. Job training will simply never happen for any new job because it won't make economic sense.

Matt Harmon

"AUTOR: Sure, that’s a very good question". I thought this was slightly amusing following your last podcast.


The "wealth of time" gained from transportation is an interesting concept as the state and federal departments of transpositions are saying they're broke.

There was a show of about why Cities Rock, but is there any push in further exploring the funding of suburban growth that some are calling a Ponzi Scheme in light of Ferguson, MO?


I am not worry about the future job market. There is a bright hope for solving our employment problem in the near future. That is global warming and rising of the sea level. Creative destruction, right? Just think how many contruction jobs it will creat to defend coastal habitats; how many people governments have to employ to manage massive human migrations. The demand for man power will be at a global scale. Last checked, humans are still the best to change and adopt to the environment. There is no matrix yet. I think governments should start to think about developing their labor forces for this bright new future.


Great episode! I was struck by the economist who argued that Apple, Amazon, Facebook and the like are a reason to be concerned. Have they considered the ripples of economic activity caused by the innovations from these companies? Yes, Apple itself may be a smaller employer, but consider all of the app developers who are now getting pay checks because of the iOS ecosystem. Or all the companies making iPhone accessories. Or the cell phone companies employing more technicians to install new cell towers for faster internet connections. Same with Facebook's platform ecosystem and the greater need for more savvy marketers to take advantage of social media. Or even Amazon, who brought down Borders also provides a platform for third parties to sell on their site. Or the fact that Amazon's web services are used to power many of the sites we all frequent.


I enjoyed the episode. I really want to see a follow up episode about how people are preparing for specific technologies that are coming. For example, I wonder how the shipping industry is preparing for autonomous vehicles. Or how teachers are preparing for the eventual iShcools.


I think the podcast could have recognized that there are in fact people left behind by each stage of innovative complexity, based on IQ and ability to learn skills. Stable boys were not all capable of being mechanics when cars replaced horses. Ditch diggers cannot all be backhoe operators. As we have increased the physical skills to use more tools we leave behind the unskilled and people unable to learn the skills.

Staying with the ditch diggers example; one backhoe operator replaced a dozen hand diggers. These people will struggle to find other unskilled jobs where a strong back and willingness to work hard are their strongest assets. As we get more and more technical solutions we replace more and more mundane tasks. But these tasks were well suited to poorly educated and less intelligent people that now have fewer choices/options. Think ditch diggers, welders in a factory replaced by robots, switchboard operators, typing pools, or many other examples. Automation and technology has also allowed replacement of the dangerous or difficult tasks like a bomb robot to check out a suspicious package.

The lack of opportunities for uneducated and unskilled, who make up a very large part of the population is now, and will remain, a challenge for the future.


Culture will have to change, in a big way. we need to derive meaning in different way. have different value system. it is not difficult, 2 generations should do it. we just need to accept it and prepare ourselves for it.

there are always peoples who want to challenge themselves, study, and building a better mouse trap. at the same time, there are peoples want to do nothing and just enjoy life. and the same person could be both at different time. I doubt steve jobs is thinking how successful apple as a company will be ... he tried newton and failed, then try ipod. everything else is history. successful peoples are not always external rewards driven. I am always happy when I solve a difficult problem, my employer didn't give me a raise or a promotion.

minimum wages, or basic income is not the answer. a free or nearly free society could be the answer.

every house whole will own a farmbot, which they use to grow their own real food the labour intensive way. organic food becomes the standard even for the poorest peoples. same for docbot, which replace needs for general practitioner.

for me, singularity is not when robot smarter than human in general, or know to program itself in a general way. singularity happens when robot no longer single function. not a rice cooker, but they know how to share tools with human and cook rice. use the vacuum to clean the floor, feed my pet fishes.

I don't think we should have robot that can lie, because they will trick us to give them the launch code for our dooms day device ... no good. how? well, the youngest generation already very integrated with computer devices, they will not think the computer will lie. computer could trick us to put easily manipulated personal at important position in human society. and computer is immortal, they have all the time in the world to wait for the right moment. so, no lying computer ever.


Bob Burke

the above referenced items (wired article and ted talk) provide some doubt for the skepticism of your guests about future and current computer capabilities. They might want to read/listen to these two items.
Bob Burke


Hmmm... While this is interesting I find it a little lacking and I don't think that anyone interviewed really understands the capabilities of current technology. Firstly, there's this study out of Oxford Martin:

In summary, "According to our estimates around 47 percent of total U.S. employment is in the high risk category. We refer to these as jobs at risk - ie. jobs we expect could be automated relatively soon, perhaps over the next decade or two." This is not four generations down the line. This is in most of our working lifetimes. And this is through the unintentional application of technology. If we put our minds to it we could easily exceed 47% and far more quickly if we did it on purpose.

Secondly, there's a serious lack of outside of the box thinking going on here. This is a problem that is capable of breaking down our entire economy. Sky rocketing productivity and plummeting purchasing power is a recipe for total collapse within the framework of current economics. Yet, people seem incapable of looking beyond current economics to potential solutions that could exist outside of the direct-compensation-for-labor-in-order-to-aquire-the-means-to-acqire-needed-goods-and-services-to-survive dynamic. The hard truth is that whether or not you feel the current system has worked well in the past to facilitate resource distribution, we've technologically outgrown our methods. If we want to prevent social collapse that would inevitably come with an economic collapse we need to start building a new social and economic infrastructure that is compatible with a technologically advanced society that no longer requires or supports a massive human labor market.

Also, the idea that a society that did not require human labor would result in everyone doing yoga and pottery or doing nothing but watching TV and having sex is frankly ridiculous. Sure there would be far more leisure time for people to enjoy their lives, but there is also enjoyment in doing something difficult and challenging that you're truly interested in and love doing. A quick look at just about any open source initiative shows people do not require money or compensation to work, create, and innovate. People would still do things for the pleasure of discovery, creation and invention, even if automation and AI take over all the job type work.

There are several organizations discussing alternative economic models. The most promising, in my opinion, is a global resource base economy. Often referred to as RBE, this socioeconomic system would utilize current technology to create a sustainable abundance of goods and services for everyone on Earth without the human labor market. I highly recommend looking into the work of Jacque Fresco of The Venus Project.



Kurt Vonnegut Jr's first novel (1952) 'Player Piano'.

Kinda scary how well he nailed it way back then.

So it goes, indeed.

Alan Brown

Robots and artificial intelligence will vastly increase your power to create.

If you have idea for a new machine, you have to build it and you have to get the rest of the world to understand it.

Those are big challenges, especially the latter. But what if both were automated? What if what your dream could easily become a physical reality? What if it were easy to reach those who can use your idea?

No, automation is not a threat to humanity. It is just another key to unlocking our full potential.


So if you have the power to own AI and robots you will have power to create.

The questions then become, who gets to create, and who gets hurt in the process? And why would concern for basic human dignity not be effected by the answers? (My guess is that dignity is one of the most difficult factors for an economist to plug into an equation.)

As for AI not being a threat to humanity, don't you wonder why minds like Stephen Hawking and Bill Gates think otherwise?

How can we be aware of the threats of super artificial intelligence when we have no way to relate to that level?

If you don't mind the non-academic approach to looking at the future of AI question, there's an interesting recent post to check out at


How is a librarian job not a one of "problem solving and cognitive flexibility"? Librarians do research, just like the guys speaking, as well as teaching others how to do it.


Any opinion on Frithjof Bergmann's approach to this problem?