A CRITIQUE OF THE SECOND MACHINE AGE (Or the Need to Shed our Romantic Ideas about Wage Labor)

(This post is based on episode 11 of the Review the Future Podcast. For a more detailed treatment of this topic you can subscribe to the podcast via iTunes or download it from reviewthefuture.com.)

What is this Book?

The Second Machine Age is a book by Erik Brynjolfsson and Andrew McAfee that explores the impacts of new technologies on the economy. For those who are familiar with such topics, it’s not likely this book will teach you much you don’t already know. However, for the layperson, this book is an extremely well written and clear introduction to the economic pros and cons of our current digital revolution. Because of the skillful way it stitches everything together, Second Machine Age has a good chance of being one of the most important nonfiction books of 2014.

The Goal of this Blog Post

On the whole, we like The Second Machine Age. We think it tells a plausible story and for the most part we agree with its perspective. However, we have criticisms of one of the book’s later chapters, the one entitled “Long-Term Recommendations.” Thus the primary goal of this article is to articulate those criticisms. But first, for the sake of background, we will summarize some of the book’s main arguments.

A Quick Summary of Second Machine Age

According to Brynjolfsson and McAfee, exponential gains in computing, digitization of goods, and recombinant innovation are all driving rapid technological growth. Technology has begun to perform advanced mental tasks—like driving cars and understanding human speech—that were previously thought impossible. And in economic terms, these new technologies, according to the authors, are increasing both the ‘bounty’ and the ‘spread.’

Bounty is a blanket term for all of the productivity and quality of life gains provided by new technologies. Brynjolfsson and McAfee feel that the bounty of technology is growing tremendously, but, because of the limitations of our economic measures, we have a tendency to greatly underestimate the progress we are making.

Spread is a euphemism for inequality. According to the authors, technology is increasing spread because of (a) skill biased technical change, (b) capital’s increasing market share relative to labor, and (c) superstar economics. All three of these trends have some evidence backing them up, and the supposition that technology is the primary driver of these trends makes a great deal of sense.

The authors also suggest that technological unemployment—a phenomenon long thought of as impossible by mainstream economists—is in fact possible. They discuss three arguments for how technological unemployment could occur:

  1. In industries subject to inelastic demand, automation can lower the price of goods without creating any additional demand for those goods (and thus labor to make those goods). Over the long term, as human needs become relatively more satiated, this inelasticity could even apply to the economy as a whole. Such an outcome would directly undermine the luddite fallacy, which is the argument economists traditionally use to dismiss technological unemployment.
  2. If technological change is fast enough, it could outpace the speed at which workers are able to retrain and find new jobs, thereby turning short term frictional unemployment into long term structural unemployment.
  3. There is a floor on how low wages can go. If automation technology continues to drive wages down, those wages could cross a threshold below which the arrangement is not worth the employee’s time. Eventually the value of certain workers could fall so low that they are not worth hiring, even at zero price.

Policy Recommendations

The book makes several short term policy recommendations. We will not list them all here, as they represent a suite of largely uncontroversial proposals designed to speed up innovation and growth. These proposals, if they were enacted, would conceivably help to get our economy working more efficiently and increase our ability to match workers to the jobs that still need doing. They would also grow the technological bounty that makes all of our lives better. It’s hard not to agree with most of these proposals.

However, if we accept the premise that  “thinking” machines will encroach further and further into the domain of human skills, and that over the long term we are destined for not just rampant inequality but also wide scale technological unemployment, then all of the short term proposals provided by this book could actually accelerate unemployment. After all, more innovation means more and better machines, which ultimately could mean more displaced labor.

Long Term Recommendations

In this chapter, the authors address long-term concerns. In a near future where androids potentially substitute for most human labor, will the standard economics playbook still work?

Brynjolfsson and McAfee are clear about two major preferences:

  1. They do not want to slow or halt technological progress.
  2. They do not want to turn away from capitalism, by which they mean, “a decentralized economic system of production and exchange in which most of the means of production are in private hands (as opposed to belonging to the government), where most exchange is voluntary (no one can force you to sign a contract against your will), and where most goods have prices that vary based on relative supply and demand instead of being fixed by a central authority.”

We agree with these two premises. So far, so good.

Should we Adopt a Basic Income?

The authors go on to discuss an old idea: the basic income. This is a potential solution to the failure mode of capitalism known as technological unemployment. If an increasingly large number of people can longer find gainful employment, then the simplest solution might be to just pay everyone a basic income. This income would be given out to everyone in the country regardless of their circumstances. Thus it would be universal and unconditional. Such a basic income would ensure that everyone has a minimum standard of living. People would still be free to pursue additional income in the marketplace, and capitalism would proceed as usual.

Brynjolfsson and McAfee do a quick survey of all of the varied thinkers, both conservative and liberal, who have supported this idea in the past. Here’s a short list of people who favored a basic income:

Thomas Paine, Bertrand Russell, Martin Luther King Jr., James Tobin, Paul Samuelson, John Kenneth Galbraith, Milton Friedman, Friedrich Hayek, Richard Nixon

With such wide ranging endorsement for the idea of a basic income, one might expect Brynjolfsson and McAfee to jump on the bandwagon and endorse the idea themselves.

But, no! Basic income is apparently “not their first choice.” Why?

Because work, they argue, is fundamentally important to the mental wellbeing of people.  If we adopted a basic income, people might not be adequately incentivized to work. And therefore people and society would suffer on some deep psychological level.

To support this idea, Brynjolfsson and McAfee field a series of arguments.

Argument One: A Quote From Voltaire

The french enlightenment philosopher Voltaire once said, “Work saves a man from three great evils: boredom, vice, and need.” Now, Voltaire was a pretty smart guy, but whether someone from the eighteenth century has anything helpful to say about today’s technological reality seems doubtful. But for the sake of argument, let’s go ahead and examine this quote.

First of all, we’re not sure what Voltaire meant by “work.” Work can mean a lot of things. Work, in the broadest sense, could mean activities you do to upkeep your life, such as cleaning your bathroom or going grocery shopping. It could also consist of amateur hobbies that you undertake for fun, such as writing overly long blog posts.

However, this is not the definition of work that Brynjolfsson and McAfee are implying. They are implying a much more narrow definition of work as ‘wage labor’—meaning work done to serve the needs of the marketplace. Wage labor is work you do, at least in part, to earn money, so that you can continue to survive and exist in this modern world.

So let’s rephrase the quote to: “Wage labor saves a man from three great evils: boredom, vice, and need.”

Already this should start to sound a little bankrupt. Wage labor saves a man from boredom? Sure, a good job can relieve boredom. But a bad job can be one of the single biggest causes of boredom in a person’s life. We don’t have any statistics on this, but anecdotally we happen to know a lot of people who don’t particularly enjoy their jobs. And boredom is one of the biggest complaints these people have. A quick survey of the popular culture surrounding work would seem to imply that this is not a unique sentiment. We have a feeling that you, the reader—if you try hard enough—can think of at least one person who gets bored at their job.

(ADDITION: Gallup Poll Shows Thirteen Percent of Workers are Disengaged at Work)

So what about ‘vice?’ What even constitutes vice in 2014? Things you do for pleasure that are bad for you? Honestly, the word vice seems a bit anachronistic in this day and age, but we can think of some candidates for vice that are actually encouraged by wage labor:

  1. Aimless web browsing and perusing of “trash” media to ease the boredom of being stuck in a cubicle
  2. Sitting in a chair all day not exercising and slowly harming your health
  3. Drinking copious amounts of soda and coffee in order to stay awake during the hours demanded by your job
  4. Cooking less and eating more junk food because wage labor takes up too much of your time
  5. Needing a drink the second you get home in order to unwind after a stressful day of wage labor

Third on Voltaire’s list is ‘need’. But if wage labor could take care of need, we wouldn’t be having this conversation in the first place, right? Since we are speculating about a future where automation makes most work obsolete, then it is clear that in such a future most people will not be able to find lucrative wage labor. So looking ahead, wage labor cannot necessarily save a man from need any more than it can save a man from boredom or vice.

Argument Two: Autonomy, Mastery, and Purpose

Brynjolfsson and McAfee attempt to use Daniel Pink’s book Drive to further their point. Drive discusses three key motivations—autonomy, mastery and purpose—that improve performance on creative tasks. However, the authors of Second Machine Age seem to imply that (1) these qualities are needed for psychological wellbeing and (2) these qualities can best be obtained from wage labor. This is a misapplication of Drive’s actual thesis.

The three motivations described—autonomy, mastery and purpose—are not fundamental qualities of wage labor. In fact, wage labor is historically very bad at providing them. Thus, Pink’s book explains how modern businesses can specially incorporate these techniques in order to try to get better results from their workers.

Such mind hacking aside, wage labor has no special claims to autonomy, mastery, and purpose. Wage labor removes autonomy by forcing people to focus their energies on what the market thinks is important, rather than on what they themselves think is important. Mastery can just as easily be found in education, games, and hobbies. And purpose can be found in religion, philosophy, community service, family, country, your favorite sports team, or really just about anywhere.

Argument Three: Work is Tied to Self-Worth

The authors cite the work of economist Andrew Oswald who found “that joblessness lasting six months or longer harms feelings of well-being and other measures of mental health about as much as the death of a spouse, and that little of this decline is due to the loss of income; instead, it arises from a loss of self-worth.”

We don’t doubt that a loss of self-worth is a major factor contributing to the unhappiness of the long-term unemployed. However, we believe this outcome is culturally and not psychologically determined. The cultural expectations in America are that you are supposed to get a wage labor job and earn your living every day, otherwise you are seen as a freeloader, a layabout, a good-for-nothing. Jobs are seen as the premiere source of personal identity, and the first question out of most people’s mouths when they meet someone new is “what do you do?” We don’t see why these cultural expectations can’t change and in fact, if the premise of technological unemployment is correct, then they will have to change.

Laziness and doing nothing may always be looked down upon. But there is a big difference between doing nothing and being unemployed. As has already been articulated, there are many productive ways to spend one’s time that have nothing to do with wage labor. If our society fails to recognize the value of these non-wage labor pursuits, then the problem lies with society.

Today unemployment may be higher than we like, but work is still abundant enough that such a cultural expectation can remain unchallenged. But if the future looks like the one implied by Second Machine Age—a future where more and more people will be unable to find wage labor—then long-term unemployment will need to become not just normalized, but accepted. By reaffirming the importance of wage labor, Brynjolfsson and McAfee are helping to perpetuate the same social force that already makes unemployed people feel depressed and worthless.

Argument Four: Without Work Everything Goes Wrong

The authors cite studies by sociologist William Julius Wilson and social research Charles Murray that suggest unemployed people have higher proclivities towards crime, less successful marriages, and other problems that go beyond just low income.

Unlike Drive, we have not personally looked at this research so we cannot speak directly to the experimental rigor of these studies. Isolating for the effect of joblessness in real world communities is extremely difficult and requires controlling for a wide variety of complicating factors. In the case of Murray’s work, the authors seem to acknowledge this concern directly when they write “the disappearance of work was not the only force driving [the two communities] apart —Murray himself focuses on other factors—but we believe it is a very important one.”

As long as wage labor is directly tied to income, how can we be sure that what these studies are actually measuring is not “incomelessness?” In order to sidestep this issue, we would maybe like to see a study of two groups—one that receives a comfortable income without working, and one that receives an equivalent amount of money, but must work for it. What differences would exist between these two groups? Would the non-working group become aimless and depressed? Or would they simply repurpose their free time towards other productive tasks?

Negative Income Tax

After all this discussion of the fundamental importance of wage labor, one might expect Brynjolfsson and McAfee to recommend the creation of a Works Progress Administration or some other mechanism for artificially creating jobs. Instead they just double back and return to the basic income idea, only by another name.

The authors support Milton Friedman’s idea of a negative income tax. They claim that a negative income tax better incentivizes work. However, this distinction between a basic income and a negative income tax does not actually exist. Both a basic income and a negative income tax have two key features in common: they set an income floor below which people cannot fall, and at the same time they allow people to increase their relative income through labor. Thus we see no basis for the notion that a negative income tax better incentivizes work.

After doing some light research into Milton Friedman’s original statements we realized one possible source of the confusion. In this video, Friedman articulates the argument that a negative income tax will do a better job of incentivizing work than a “gap-filling” version of the basic income. This is certainly true. A gap-filling basic income would probably be a bad idea and have the problem of disincentivizing labor below a certain threshold. However, to our knowledge, none of the modern day basic income proposals are built around this gap-filling principle, so Brynjolfsson and McAfee’s distinction seen in this light would be a bit of a straw man argument.

What are the Goals?

We should not forget that wage labor is not the goal in itself. The real goals of our economy ought to be (1) alleviate people’s suffering and (2) increase the bounty through innovation. Although there are challenges involved, a basic income would seem to be a promising way to address both of these goals.

A basic income puts a floor on poverty and does so in a way that is both much simpler than our current alphabet soup of social programs, and more encouraging of autonomy. Rather than providing people with prescribed social services, people could spend their basic income dollars on whatever they feel they need. A basic income decentralizes decision making and puts the power in the hands of individuals.

As a corollary, a basic income might help unlock innovation by bringing people up to the subsistence level and thereby ensuring that they have the opportunity to compete and innovate in the market economy. Moreover, the safety net of basic income might spur entrepreneurship by reducing the risk of starting a small business. Is it possible more people would attempt to start businesses if they knew they had a cushion of basic income to protect them in the event of failure? (And as we all know, most new businesses have a high chance of failure.)

Under a basic income, there is no doubt that some people would choose to forgo wage labor altogether and live at the poverty line. But is this such a bad thing? These people would be making a personal choice. And we imagine many such people would find interesting and productive ways of spending their time that might be culturally valuable, even if they do not carry a price in the marketplace. If a musician chooses to live off of a basic income and make music, he doesn’t make money in the economy, but we all still get to enjoy his music. If a free software programmer chooses to live off a basic income, he doesn’t make money in the economy, but we all still get to enjoy his free software. If a history enthusiast chooses to live off a basic income, he doesn’t make money in the economy, but we all still get to enjoy his Wikipedia articles. As Brynjolfsson and McAfee argue earlier in the book, the value generated by digital content is not always well measured or compensated by the marketplace, but that doesn’t mean such content doesn’t improve our lives.

However, we may be preaching to the choir since Brynjolfsson and McAfee, despite their protestations, do in fact support a basic income. They just prefer the particular version of basic income that goes by the name “negative income tax.”

Pause for Skepticism

Now, it is worth noting that the “end of work” scenario is not a foregone conclusion. Here are two potential defeaters to this outcome:

  1. Human capabilities are not necessarily fixed. One byproduct of future technologies might be a redefinition of what it is to be human. If we begin to “upgrade” humans, whether through genetics or brain-computer interfaces or some other means, many technological unemployment concerns could become irrelevant. Upgradeable humans could solve both the retraining problem (just download a new skill set to your brain, matrix-style) and the issue of inelastic demand (super-humans might develop brand new classes of needs).
  2. A wide range of intangible goods—such as attention, experiences, potential, belonging, and status—might remain scarce indefinitely and continue to drive a market for human labor, even after the androids have arrived. Although it’s hard to imagine a market in such goods replacing our current manufacturing and service economy, it must have been equally hard for pre-industrial people working on farms to imagine the economy of today. Thus we may simply be lacking imagination when it comes to envisioning the jobs of the future. (For a more detailed discussion of this topic see episode 10 of the Review the Future podcast.)

Despite these defeaters, we definitely think the technological unemployment scenario is worth thinking about. First of all, the issue of timing is paramount, and at present it seems like we have a good chance of automating away many jobs long before we figure out how to upgrade human minds or develop brand new uses for human labor. Second, it won’t take anything close to full unemployment to create problems for our system. Even a twenty percent unemployment rate, (or an equivalent drop in Labor Force Participation) for example, might be enough to trigger a consumer collapse or at least great suffering and social unrest among lower classes.

Final Thought

Wage labor is a means to an end, not an end in itself. While the Second Machine Age paints a clear picture of some of the potential problems facing our economy, it fails to fully take to heart this fundamental distinction.

Self-Driving Cars Need Their Own Speed Limits

As the video above demonstrates dramatically, speed limits are a well-meaning regulation that’s going to have to be rethought in the era of self-driving cars. Eric Schmidt’s on record saying that a major problem with the current design of the Google self-driving car is that it obeys speed limits. But a computer that can safely drive can do so at faster speeds than a human. Further, a car specially designed from the ground up to be piloted by machine could have acceleration and braking systems optimized for the machine’s much faster reaction time, so cars that use self-driving technology might become rapidly faster and more efficient.

We need to establish an objective procedure, by which autonomous vehicles can demonstrate safety at various speeds, and if we are going to continue to have speed limits (I think we should!) they should be designed to maximize safety and efficiency while keeping up with current-generation technology.

(H/t +Wayne Radinsky for the video)

Some Possible Effects of Near Future Technologies on Real Estate Prices

Real estate is one of those goods that will remain scarce for the foreseeable future. It is hard to imagine a time when territory on planet Earth will not still be considered a valuable resource.

However, real estate prices might be dramatically affected (and in many cases lowered) by some near future technologies that are already on their way:

  • Automated construction techniques will allow buildings to be fabricated with less workers, at a lower cost, and in less time. Thus the price of actual structures, if not necessarily the underlying property, should fall accordingly.
  • Likewise, new automation-enabled architectural designs might allow the creation of structures that more efficiently and comfortably fit larger numbers of people within the same plot of land.
  • The growing efficacy and acceptance of telecommunication technologies should allow increasing numbers of people to choose where they live regardless of whether that location is near a job site.
  • Growing numbers of permanently unemployed people who have given up and dropped out of the labor force may find it increasingly compelling to move away from cities and other centers of economic activity. If they are not going to find work anyway, they might as well live where things are cheaper.
  • Commuting in self-driving cars will be a vastly improved experience—why not enjoy a movie or a nap during your two hour drive? Thus it may become increasingly viable to live farther away not just from your job site but from your loved ones and other amenities. Do you necessarily need to live right next to nightlife, for example, if an automated chauffeur will drop you off wherever you want and then pick you up at the end of the night?
  • Virtual reality will improve our currently narrow-bandwidth communications technologies. Over time, people will be able to get more and more of the benefits of being “face to face” by simply connecting remotely.
  • Eventually virtual reality may even get good enough that people will be able to tolerate living in much cheaper and more cramped living quarters. What if after you got home, you could put on a headset and be transported to a virtual mansion?
  • New food production methods (such as lab-grown meat) might allow us to reclaim land currently devoted to tasks like farming. Similarly, technological disruption of industries might lead to many formerly commercial districts getting repurposed as residential.

Erik Brynjolfsson Diagnoses the Problem in the Economy But Has No Solution

In this talk Erik Brynjolfsson clearly makes the case that productivity and employment are decoupling from each other. His presentation is a fantastic description of what is happening today and a fitting answer to the stagnationists.

That said, his solution at the end of this video amounts to little more than a clever turn of phrase: namely he suggests that we need to race with machines. In my detailed review of Erik’s book Race Against the Machine I criticized this idea:

The first suggestion the authors make can be summarized as “race with machines.” A human-machine combo has the potential to be much more powerful than either a human or machine alone. So therefore it’s not simply a question of machines replacing humans. It’s a question of how can humans and machines best work together.

I don’t disagree with this point on the surface. But I fail to see how it suggests a way out of our current predicament. The human-machine combo is a major cause of the superstar economics described earlier in the book. Strengthen the human-machine combo and the superstar effect will only get worse. In addition, if computers are encroaching further and further into the world of human skills, won’t the percentage of human in the human-machine partnership just keep shrinking? And at an exponential pace?

Moreover, as I’ve written about before on this site, the human-machine partnership can sometimes be less than the sum of its parts. Consider the example of airline pilots:

“In a draft report cited by the Associated Press in July, the agency stated that pilots sometimes “abdicate too much responsibility to automated systems.” Automation encumbers pilots with too much help, and at some point the babysitter becomes the baby, hindering the software rather than helping it. This is the problem of “de-skilling,” and it is an argument for either using humans alone, or machines alone, but not putting them together.”

At some point it may be possible to literally race with machines in the sense of actually merging man and machine together. But this has not been the current trend. What we have been seeing instead is people offloading cognitive tasks to independent machine algorithms. How many of us remember phone numbers anymore? Indeed memory has been one of the first cognitive tasks to get offloaded.

In order to race with machines I am convinced we need to actually enhance human intelligence directly. This is probably not impossible, but will require a much better understanding of the brain, and as a solution it will probably not arrive in time to stave off the massive decoupling that is affecting our economy.

Here is Eliezer Yudkowsky on the relative difficulty of agumenting humans versus developing standalone artificial intelligence:

“I originally gave the example of humans augmented with brain-computer interfaces, using their improved intelligence to build better brain-computer interfaces. A difficulty with this scenario is that there’s two parts to the system, the brain and the computer. If you want to improve the complete system, you can build interfaces with higher neural bandwidth to more powerful computers that do more cognitive work. But sooner or later you run into a bottleneck, which is the brain part of the brain-computer system. The core of your system has a serial speed of around a hundred operations per second. And worse, you can’t reprogram it. Evolution did not build human brains to be hacked. Even if on the hardware level we could read and modify each individual neuron, and add neurons, and speed up neurons, we’d still be in trouble because the brain’s software is a huge mess of undocumented spaghetti code. The human brain is not end-user-modifiable.

“So trying to use brain-computer interfaces to create smarter-than-human intelligence may be like trying to build a heavier-than-air flying machine by strapping jet engines onto a bird. I’m not saying it could never, ever be done. But we might need a smarter-than-human AI just to handle the job of upgrading humans, especially if we want the upgrading process to be safe, sane, healthy, and pleasant. Upgrading humans may take a much more advanced technology, and a much more advanced understanding of cognitive science, than starting over and building a mind from scratch.”

Is Programming Really as Future Proof a Profession as People Think?

While the job market as a whole is troubled, in certain high tech fields, such as programming, labor demand is still quite high. But while times are good for programmers, is programming actually a future proof profession over the long haul?

One line of reasoning would suggest that yes, programmers are going to be safe in the new economy. After all, the logic goes, even if robots take all our jobs, someone still has to tell the robots what to do, and those people are programmers.

But let me suggest a different way of looking at things: A programmer is really just a translator. A programmer essentially translates a natural language idea, like “I need an app that does X” into machine-friendly code. And translation is a data processing task that computers are getting increasingly good at performing.

Imagine an extremely high-level programming language, one almost identical to natural language. You simply describe the program you want to build and the compiler handles the rest. Generally high level languages carry a performance cost, but in a future ecosystem rife with cheap computing power, such a cost might be negligible.

If that scenario seems too far-fetched, let’s try a different angle: how big is the possibility space of useful everyday programs? It certainly can’t be limitless. Remember that the goal of a good programmer is not necessarily just to write code that works, but also to write code that is modular and reusable for a wide variety of tasks. So as the library of useful code grows, is it possible that eventually most of the important everyday programming tasks will have been handled? That there will be an ever shrinking frontier of new code to write, and an ever shrinking group of programmers exploring that frontier? I’m not saying there will be no programmers. Just that after a while there might be far fewer than current demand would suggest. In other words, programming could be one of those ironic professions where doing it truly well means making yourself obsolete.

Along these lines, here’s a revealing quote from programmer Jason Lewis on his blog Practical Elegance:

“Marc Andreessen famously explained ‘Why Software Is Eating The World’ in the WSJ a couple of years ago. What he failed to mention is that the snake of software is also quietly eating its own tail.

“I’m not just an old-fashioned Job Destroyer, replacing secretaries and mid-level bureaucracy with CRM and accounting suites. By using the most efficient possible languages (Ruby and Clojure, in my case, rather than Java or C#) and relying on free and open source software (Postgres rather than Oracle, for instance), I’m potentially destroying jobs in my own sector!”

“Fan Out” is a Term For the Number of Robots Controlled By A Single Human

This idea comes from a book by Illa Reza Nourbakhsh.

Apparently “fan out” is the term for the robot to human ratio in a man-machine relationship.

“In USAR, the effective number of robots controlled by a single human operator has a formal term: fan out… Ironically, fielded robots have very low fan out scores today.  For instance, the Predator-class drones, unmanned aerial vehicles that fight proxy battles for the United States in distant lands, have a fan out of less than 0.2.  That is, more than five people are required at all times, just to manage a single robot.  In USAR, researchers have begun to demonstrate ever-increasing fan out — exceeding 6.0 — by providing the robots with more and more autonomy so that the human operator is only responsible for the most strategic decisions, with robots making every tactical choice.”

This is another variable I wouldn’t mind seeing plotted against time as a possible indicator of technological unemployment. Within the scope of a particular industry or robotic application, one might expect to see this variable steadily increase.

Tip via: Marginal Revolution

We Need to Get Watson The Fifty Percent of Clinical Trial Data That’s Missing

As you’ve probably heard, IBM is currently repurposing its Jeopardy-winning super computer, Watson, to preform medical diagnosis. One of Watson’s great advantages over human doctors is the ability to read and digest vast volumes of medical literature that no person would ever have the time to get through.

On a related topic, I recently heard an interview with Ben Goldacre which contained the following disconcerting fact:

“Half of all clinical trials for the treatments we use today have never been published, and trials with positive results are twice as likely to be published.”

As if the implications of this statistic aren’t bad enough—namely that drugs are systematically made to look more effective than they actually are due to passive censorship of failed studies—the arrival of algorithms like Watson on the scene makes the need to release all trials even more imperative. Algorithms dine on data, and the more data they consume the better they will perform. When the algorithm in question is as critical as medical diagnosis, we should be feeding it every piece of information we can possibly find. Quite literally someone’s life could depend on it.

Video Simulation of How Autonomous Cars Will Navigate Intersections

This fun little video offers an aerial simulation of how a crowd of self-driving cars might navigate a busy intersection. I had two reactions upon watching:

  1. I am excited about how efficient self-driving cars are going to be at using existing roads. I hope to spend a lot less time waiting in traffic.
  2. If you imagine yourself inside one of the vehicles in this simulation, its clear how scary riding in one of these cars might be for an first timer. The cars routinely head right towards each other only to swerve away at the last moment.

Numbers: Projected Job Losses from Driverless Cars

This is a long post with a lot of correct (if perhaps obvious) points. But I got excited about this part near the end, where there are some numbers. I’m not entirely sure where these numbers are from, but they are interesting to think about.

At the same time, driverless cars will dramatically affect employment around the world.

  • Over time over 232,000 taxi and limo drivers in the U.S. will lose their jobs.
  • Over 647,000 bus drivers will be out of work.
  • Over 125,000 truck drivers will be looking for new careers.
  • Other jobs affected will include jobs at gas stations, parking lots, car washes, traffic cops, traffic courts, doctors, nurses, pizza delivery, mail delivery, FedEx and UPS jobs, as well as vehicle manufacturing positions.

In the future, the number of vehicles sold will begin to decline.

Report Claims that Robots Create Jobs

At this site a central question of: “Is technology destroying jobs faster than it creates them,” motivates a fair amount of the discussion. Here’s a political article that links to a report that comes to the negative conclusion, at least considering robotics and manufacturing. I’d like to hear some push-back on this, if any can be made, from the data side (I’m not really interested in the politics either way). What do you think? Are they missing something important?