Countering the march of the robots narrative

I read a very interesting Report last week – False Alarmism: Technological Disruption and the U.S. Labor Market, 1850–2015 – published on May 8, 2017 by the Information Technology and Innovation Foundation (ITIF) and written by Robert Atkinson and John Wu. The title is indicative of the message. Somehow, contemporary commentators including many on the so-called progressive Left are stuck in the ‘robots are coming for your jobs’ narrative, which then somehow morphs into a resignation that there will never be enough jobs for all those who desire them, and then surrender, we need a basic income to keep people eating. Apparently, then human creativity will spring forth from the despair of unemployment because the pittance received from the basic income will allow people to engage their inner entrepreneurial spirit with businesses popping up all over the place, great works of art and music being pumped out and all the rest of the basic income camp’s vision of blithe happiness. Pigs might fly! Of course, if this was happening at the pace that some would have us believe then productivity growth would be booming and investment to GDP ratios high. The robots camp then say – well it is only a matter of time – business needs time to adapt to the new technologies available (for example, Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics). Technological change is on-going and there have been great leaps in techniques in history. But the ITIF research suggests that the current era does not signal it is one of these great leaps, and, in fact, the “US labor market is experiencing unprecedented calm” right now.

A report earlier this year (March 26, 2017) in – Tech world debate on robots and jobs heats up – aired the view of “a tech entrepreneur” who claimed that:

… some 80 to 90 percent of jobs will be eliminated in the next 10 to 15 years.

Of course, the discussion didn’t question whether the “tech entrepreneur” had a vested commercial interest in making these radical predictions. It just repeated the claims.

The Founder and Executive Chairman of the World Economic Forum is certainly pushing the ‘robots are coming’ view, which means the so-called leaders of the world are getting it rammed down their throats along with all the fine food and wine they consume at the WEF meetings.

He (Klaus Schwab) wrote on January 14, 2016 – The Fourth Industrial Revolution: what it means, how to respond – that:

We stand on the brink of a technological revolution that will fundamentally alter the way we live, work, and relate to one another. In its scale, scope, and complexity, the transformation will be unlike anything humankind has experienced before …

The speed of current breakthroughs has no historical precedent. When compared with previous industrial revolutions, the Fourth is evolving at an exponential rather than a linear pace. Moreover, it is disrupting almost every industry in every country. And the breadth and depth of these changes herald the transformation of entire systems of production, management, and governance.

The ITIF Report – False Alarmism: Technological Disruption and the U.S. Labor Market, 1850–2015 – is motivated by claims such as the above, which it calls the recent:

… article of faith that workers in advanced industrial nations face almost unprecedented levels of labor-market disruption and insecurity.

The ITIF study notes, by way of contradiction of Schwab’s assertion that “the pace of technical change has not accelerated over the last 200 years, and little evidence exists that this will change going forward”.

They suggest each historical epoch is marked by doomsayers who believe they are living through technological changes on a scale that is unprecedented.

Their report follows earlier studies by ITIF authors which cast doubt on the popular narrative:

1. Robert D. Atkinson (2017) – In Defense of Robots – National Review, April 17, 2017.

2. Ben Miller and Robert D. Atkinson (2013) – Are Robots Taking Our Jobs, or Making Them? – September 2013.

In the current Report, the ITIF researchers attempt to redress what they call the “ahistorical” nature of the “pessimists’ grim assessments”, by examining the “the last 165 years of American history” where the “statistics show that the U.S. labor market is not experiencing particularly high levels of job churn”.

They define “job churn”:

… as new occupations being created while older occupations are destroyed.

ITIF built an interesting dataset – which collated “U.S. occupational trends from 1850 to 2015” from Census data.

They computed “changes in occupational job levels for each intervening decade and through coding each occupation (for sensitivity to technological shifts) were able ” to judge whether increases or decreases in employment in a given decade were likely due to technological progress or other factors”.

They study churn because they note that technology impacts on labour markets through occupational shifts.

First, “technology changes occupational structures … through transforming products and industries.” For example, shifts in jobs associated with phonograph production to CDs and, more recent, digital music forms.

Second, “Technology can also eliminate jobs by allowing some occupations to be more productive”. For example, human pinspotters in bowling alleys being replaced by automatic machines or the “self-service elevators” in buildings.

They note that:

… while 180,000 Americans were employed as travel agents at the turn of the millennium, with the emergence of Internet-based travel booking, just over 90,000 were employed in 2015.

The list is long.

But while occupations disappear, new occupations emerge.

All of these shifts see a particular occupation grow, reach a peak and then decline rather sharply as new occupations (that replace them) evolve.

In my blog last week – Automation and full employment – back to the 1960s – I noted the emergence of a shortage of drone pilots.

Many occupations today “didn’t exist 30 years ago” and were created because “technology made them possible”.

Some occupations grow because it is “hard to improve worker productivity” – for example, “it still takes one teacher to teach 30 students in elementary school, just as it did 40 years ago”. So as population grows, so to does employment in these areas.

I won’t discuss the specific methodology used in the ITIF research. You can consult the Report if interested. Suffice to say they compile two different datasets (one using 1950 occupational classifications and the other using 2010 classifications).

The more recent classification overcomes criticisms that the 1950s classification “fails to capture some detailed occupational shifts (especially in the decades at the front end and back end of this data set)”.

The overall results are not particularly sensitive to which dataset is deployed. The “same general trend” is revealed.

To measures of “job churn” are computed:

1. “change in each occupation relative to overall occupational change … even if an occupation doesn’t lose jobs, if it didn’t grow as fast as the overall labor market, the delta between that growth and overall labor force growth would be calculated as churn … Absolute values were taken of negative numbers, and the sum of employment change was calculated for all occupations. This was then divided by the number of jobs at the beginning of the decade to measure the rate of churn.”

2. “a related measure. For declining occupations, it only includes a sum of occupational changes where there was an absolute loss in jobs … the churn rates using this method were lower than the first method … However, the correlation between these two measures of churn was quite robust (0.92 for 1950 occupation codes and 0.90 for 2010 occupation codes respectively) suggesting that either measure captures generally the same historical change”.

Concentrating on the first method of computing job churn, ITIF produced “Figure 7: Rate of Occupational Change by Decade (1950s Categories).

The findings are obvious:

Rather than increasing, the rate of occupational churn in the last few decades is the lowest in American history, at least since 1850 … occupational churn peaked at 50 percent in the decades between 1850 and 1870 … it was still above 25 percent for the decades from 1920 to 1980 … it fell to around 20 percent in the 1980s and 1990s, to just 14 percent in the 2000s, and 6 percent in the first half of the 2010s.

The next graph (ITIFs Figure 9) shows the calculations using the 2010 occupational classification which is consistent back to 1950.

The same story is revealed.

Their overall conclusions:

1. “Levels of occupational churn in the United States are now at historic lows”.

2. “The levels of churn in the last 20 years—a period of the dot-com crash, the financial crisis of 2007 to 2008, the subsequent Great Recession, and the emergence of new technologies that are purported to be more powerfully disruptive than anything in the past—have been just 38 percent of the levels from 1950 to 2000, and 42 percent of the levels from 1850 to 2000.”

3. “the single biggest economic challenge facing advanced economies today is not too much labor market churn, but too little, and thus too little productivity growth. Increasing productivity is the only way to improve living standards—yet productivity in the last decade has advanced at the slowest rate in 60 years.”

They also point out that when Klaus Schwab predicts that “robotic and artificial intelligence will destroy 5 millino jobs by 2020”, which sounds like a lot of jobs, in reality, “it adds up to the elimination of just 0.25 percent of jobs annually for next five years … barely a rounding error”.

They critique the 2013 Oxford study – The Future of Employment: How Susceptible are Jobs to Computerisation – that is being used by the ‘robots are coming’ camp.

They note it was never submitted to peer review and used a flawed methodology that failed to “examine all 702 U.S. occupational categories”. When more careful work is done on this basis, the Oxford figure that 47 percent of U.S. jobs will go in the next two decades, becomes more like 10 per cent.

Why does all this matter (according to ITIF)?

They say this is important because the popular narrative has led to policies that would be destructive being proposed and considered – such as Bill Gates robot tax. They say that particular policy would be like “taxing tractors in the 1920s”.


The message of the study is threefold:

1. The hysteria surrounding the ‘robots are coming’ narrative is just that hysteria. They advise everyone to “Take a deep breath, and calm down.”

2. The risk is “that technological change and resulting productivity growth will be too slow, not too fast”.

3. “Policymakers should do more to improve labor-market transitions for workers who lose their jobs … it doesn’t matter whether the losses stem from short-term business-cycle downturns or from trends that lead to natural labor-market churn”.

Which is where the Job Guarantee and UBI debate is centred. What do we do with workers that lose their jobs in recession and/or via churn?

The first logical question is do the workers want to keep working? That means the state needs to create jobs to shore up income losses not just rely on income transfers.

That will be the best solution for almost all workers displaced by technology who value work.

The series so far

This is a further part of a series I am writing as background to my next book with Joan Muysken analysing the Future of Work. More instalments will come as the research process unfolds.

The series so far:

1. When Austrians ate dogs.

2. Employment as a human right.

3. The rise of the “private government.

4. The evolution of full employment legislation in the US.

5. Automation and full employment – back to the 1960s.

6. Countering the march of the robots narrative.

The blogs in these series should be considered working notes rather than self-contained topics. Ultimately, they will be edited into the final manuscript of my next book due in 2018. The book will likely be published by Edward Elgar (UK).

That is enough for today!

(c) Copyright 2017 William Mitchell. All Rights Reserved..

Spread the word ...
    This entry was posted in Demise of the Left, Future of Work. Bookmark the permalink.

    30 Responses to Countering the march of the robots narrative

    1. Alan Longbon says:

      Mention of the Gates robot tax suggestion reminds of the equally inspired Dickensian English window tax. The window tax led to windows being bricked over, a lack of daylight inside dwellings and then a generation of children blighted by Ricketts. Ricketts is a disease caused by a lack of sunlight exposure.

      A robot tax would lead to less robots and a generation blighted by stunted productivity growth, lower incomes, lower wealth and therefore lower health and education outcomes.

    2. Francisco J Flores says:

      Displaced workers means more people available to do other stuff. In any event, may we suggest a Job Gty as part of a Full Employment Fiscal Policy? Here’s how:

    3. GLH says:

      I would like to predict that in the next 10 years or decade, whichever comes first, that some 80 or 90 percent of tech entrepreneur prophets will disappear.

    4. dnm says:

      Lots of very interesting data in that report. Too much to think about to make an intellegent comment. So here’s a superficial one instead:

      What is your guess for category with the biggest absolute job losses in the period 2010-2015?

      If you guessed “the military” (79% employment loss), then pat yourself on the back. BTW this employment category is not rated as technology-affected, so I guess the authors haven’t heard the news about the drone situation. I guess they are putting the reduction down to the peace dividend.

    5. /lars says:

      Basic reason for any production is to fulfill human wants and needs. Will robots put an end to human ingenuity to create human wants and needs?
      Anyhow in most parts of the world the part of the human lifespan is occupied to wage labor and necessary education for this is very large. If the necessary time spent on wage-labor would half it would be a great achievement, could the robots deliver on that?

    6. Sandra Crawford says:

      Off topic but not the blog. An article to delight in the Guardian this morning?

      Phil McDuff
      Getting rid of deficits is disastrous for economies!! Quotes Stephanie Kelton.

    7. Gogs says:

      Under a Sunday Times headline yesterday entitled “Look up: financial red lights are flashing again”, Niall Ferguson considered several reasons to be nervous of the future. He referred to a “demographic inflection point” in which the ratio of workers to consumers has peaked; China’s working age population for instance is projected to shrink dramatically.

      On July 21st 2010 an earlier Niall Ferguson argument was demolished under Bill’s blog headline “The myth of rational expectations”.

      I wonder what he he would make of Ferguson’s latest prophecies.

    8. /lars says:

      US employment rate have been in overall decline since about 2000 when Clintons private debt bubble popped. Minor recovery 04-08 and mostly flat since 2010. For US that have fairly little programs that artificially lift people into employment rate this is probably the better indicator of the state of the labor market than basic unemployment rate.

    9. Nigel Hargreaves says:

      Apropos the blog, the one (only 1) thing I agreed with in Philip Hammond’s interview with Andrew Marr yesterday was that robots are not throwing people out of work. Unfortunately, he slipped up by saying there are no unemployed and Marr immediately leapt on that by pointing out there are 1.2 million unemployed.

      Also, I recently read a book “The second Machine Age” by Erik Brynjolfsson and Andrew McAfee which also shows that robots are not going replace humans any time soon. And BTW they advocate JG rather than UBI.

    10. Sandra Crawford says:

      I got upset with Paul Mason at the Brighton Labour Party conference this year because he gave a talk in which he showed concern for the technological advances that would, he thought reduce jobs.

      He is a progressive journalist so I would have expected something better.

      The whole point of work and the economy for that matter is to help human existence by providing it with a means to live in a civilised society. A job is a substitute for owning and tilling the land, which became impossible for everyone to do after the enclosures in Britain. We need an occupation to keep our minds healthy, and to socialise, and a living to pay our bills. Whether we have machines at work or not is irrelevant, we still need to be a part of it.

      Obviously the global oligarchs do not see the economy that way, they only look at the bottom line. But they have been doing that forever – with or without technology, in the form of off-shoring jobs to China where working conditions are poor and cheap, and centuries before that in the form of slavery. In the USA, the slavery is private prisons. A free lunch for the rich takes many guises, of which technology has become one type.

    11. Nigel Hargreaves says:

      The Guardian article is great for those of us that understand how the monetary system works, but for anyone that has been brainwashed by the neoliberal propaganda (that means almost everyone) it is worse than useless. You’ve only got to look at the comments, of which there are scores, to see that only a very small number agree with the author. Most say it is left-wing nonsense. Zimbabwe, Greece et all appear. You can’t make people understand with a few dozen words, so it is better not to say anything.

    12. Sandra Crawford says:

      Nigel – at least the article itself is fairly sound, and it is very rare to see it in the mainstream press. We have got to start somewhere. There will always be trolls refuting articles that do not follow the neoliberal mantra. That should not put anyone off or we will get nowhere.

      Perhaps you should put in some helpful links below the line!

    13. Simon Cohen says:

      Sandra, Nigel,

      The article is indeed a good one and a bright and cheery thing for anyone interested in the reality of the Monetary system and how its misunderstanding protects vested interests.

      Steph Kelton is brilliant at posing simple but pertinent questions as a way of getting the basic ideas across, she is an excellent educator in this way, for example:

      “Now,” she would say, “suppose you could eliminate all the US Treasury securities from existence. Would you do that?” She recalls this would often result in more hedging and nuanced response as the senators mulled over the implications of such a move. “But,” she would point out, “you just answered that question. You said you wanted to get rid of the debt. The treasuries are the debt.”

      That’s a great way of stumping a neo-liberal and getting across that the deficit IS the wealth of the private sector.

      people want to hold onto their view that public debt is bad because they seem to hold it like a religious belief and their universe collapses if they can’t. More tricky is the psychology of letting go of this belief which would reveal that we’ve been subjected to a 40 year scam – perhaps many could just not swallow that?

    14. Lizette says:

      It makes me think about the poem C. P. Cavafy wrote called “Waiting for the Barbarians”

      Or, what Alfred North Whitehead would call the fallacy of misplaced concreteness, also known as reification concretism,or hypostatization) is a fallacy of ambiguity, when an abstraction (abstract belief or hypothetical construct) is treated as if it were a concrete real event or physical entity. In other words, it is the error of treating something that is not concrete, such as an idea, as a concrete thing.

      Want some concreteness? Watch this and start imagining what it would take to accomplish this robot scenario \_(ヅ)_/

    15. Paulo Marques says:


      Short time reader, first time poster… Which just means that I’m new here and that your ideas resonated with me (and reality). I still don’t know how practical matters limit MTT, but I’m sure you’ve written about it and I’ll get to it at some point.
      This is to be clear that I don’t disagree with your analysis, especially after the last post in the series. I think the fear from AI comes from the fact they it suddenly can do a lot of things that we didn’t expect it to be able to do (outside of academia). It’s a field that I’m following very closely, though I don’t think my career will go down such an interesting path, and sometimes it’s frightening to see a recent paper that will make human tasks redundant in the future.
      In the end, as you say, the issue isn’t technological: people can more easily adapt to similar tasks that are much harder for an AI before the next evolution is made, not to mention that it all requires a fair bit of limited resources and a lot of power. But still… it wouldn’t surprise me to find many specific jobs going away pretty quickly.

      I wouldn’t expect the common politician, or even the common software engineer to have a clue about how any of it works, and, yes, a lot of them are using the fad to promote their own agenda without understanding a thing; that’s nothing new or unique. But it’s pretty clear it will have a big impact reasonably quickly on repetitive tasks that need minimal supervision.

    16. Stone says:

      There is some truth both in this (very interesting -thanks!) post and also in your previous excelent post “The Rise of The Private Government”. But those two posts give me contradictory messages regarding UBI. Your earlier post harks back to an ideal of individual control over individual labour and enterprise. That is something that UBI advocates make a case for. It makes sense to me that a UBI could generate enough aggregate demand to provide custom for such individual enterprises (or co-ops formed by groups of enterprising UBI recipients etc). A UBI could actually bring on the march of the robots because workers could refuse to do tedious mundane tasks and new start-ups etc could bring on new robot technologies etc. Robot fruit pickers have been around for over a decade but low paid exploited workers have meant that there is no incentive to invest in such technology.
      By contrast, in this post you say that we all need to have bosses telling us what to do or else we will fail to acheive our potential etc. I’m particularly troubled by the labour relation situation under a job guarantee. Could JG recipients strike for better pay or conditions? The JG is supposed to be conducting work that can be expanded or reduced swiftly in response to macroeconomic needs. My understanding is that rewarding work is work that needs to be done and commands respect and perhaps even puts the worker in a strong bargaining position where refusual to work would jeopodise society etc. Can a JG really be like that?

    17. CS says:

      Very interesting article. I must admit I had become convinced of the robots/UBI narrative before discovering MMT. Now I see things very differently. The kind of people advocating UBI make me very nervous. As an underemployed precariat type person, I realise the value of productive work and as much as I enjoy spending all day reading about MMT, I wish I had more stable employment cause being alone all day tapping on a computer is not the best for the soul -even if it is reading about MMT.

    18. Gordon says:


      You say – The risk is “that technological change and resulting productivity growth will be too slow, not too fast”.

      Is one possible reason for this that companies repackage basically the same stuff, and market it as a next generation or advancement? I’m thinking that it’s cheaper to invest in marketing than in r+d, and many of the wealthy who control companies are only interested in easy profits. Our society is mostly rentier based now.

      On the subject of robots, I look forward to doing work of my choosing, that I enjoy, for an improved quality of life, while robots do the nuts and bolts jobs for society. Im sure that’s what all these tech entrepreneurs have in mind. I wonder if the robots will go on strike…

    19. Barri Mundee says:


      People on the JG could in theory go on strike for better conditions but they would get far better pay and conditions than on unemployment benefits. In Australia the minimum wage is much higher- $1390 per fortnight compared to $539 for the so-called Newstart.
      Those in the JG system can always be offered a better paying job either in the private sector or in the public sector.

      The JG has been carefully considered and designed. Just check by typing it in the search bar (top right of every blog).

    20. Some Guy says:

      Stone:By contrast, in this post you say that we all need to have bosses telling us what to do or else we will fail to acheive our potential etc.

      Your earlier post harks back to an ideal of individual control over individual labour and enterprise. That is something that UBI advocates make a case for.

      A UBI is just magical thinking at best. What the UBI advocates say and what the UBI does are very different things. A UBI puts bosses more firmly in control than ever. A JG frees people of boss’s power as much as possible. That’s why so many bosses are UBI advocates and so many bosses have always fought full employment tooth and nail. They understand things, while most UBI supporters have everything backwards.

      Think about what happens, in simple terms, in a real world UBI where things don’t magically happen just because people want them to, and you will come to the same conclusion that CS did.

    21. sam w says:

      A good litmus test for the integrity of those pushing the inevitability of the ‘march of the robots’:
      Read some of the stuff futurists whom are large proponents of AI are actually saying in true technical language. Eg: the ones actually well versed with computer science for example.
      They still tell us:
      There is still decades (projected/guessed) until you get ‘robots’ in the sense of automatons that interface with the acuity and finesse of humans (upright apes with opposable thumbs, high dexterity, and in may different situations) and neural networks which may simulate complex computation for coordination.
      Put this to test with personalities who are large proponents of automation and so/or should know what they are talking about eg: Elon Musk (whom there is a lot to like about).
      Add the media spin. The LITERALLY do liken AI TO EQUAL HUMAN INTELLIGENCE soon fact is its nowhere near that. This is where the crux of the argument is made and you see the bias.

      Musk and others NEVER call out the torrent of hype and factual misunderstanding in the media over their facts about AI or automation.

      AI is a blanket term, automation is a blanket term it covers a wide spectrum of definitions. Eg: the term AI might be used for a simple neural network represented by a few ‘if’ ‘else’ switches that recognises an output correctly >50% of the time given a valid input. Or it could be a deep neural network self annealed on a supercomupter given giant multi petabyte datasets that accomplishes the same result.
      With Automation: I can for instance argue that the robot revolution already took place 50 years ago: as soon as computerised mould injection, welders, CNC and complex automated assembly lines became necessities for manufacturers.

      This is the problem all when all these issues are then projected into the sphere of neoliberal ‘understanding’ of economics and politicised to make profit its a huge distortion of what is actually happening.

    22. Jim Green says:

      There is no reason why a JG cannot go hand in hand with UBI. In his eagerness to promote the laudable aims of the JG, Bill fails to appreciate that, for most supporters of UBI (in the UK, at least), UBI is seen not as a replacement for work, but as a replacement for our failing and punitive welfare system.

    23. Leftwinghillbillyprospector says:

      Robots may not take everyone’s job any time soon but I feel they will make a negative impact regardless – you do not need to actually replace your human workforce with automation, you merely need to demonstrate to them that you could do so. So forget a pay rise and in fact you’d better take a pay cut or you will be replaced by a bot.

      I recently saw a bricklaying robot that could do in an hour what it currently takes two skilled human brickies all day to do – this sort of thing is a powerful industrial weapon in the hands of capital.

      I think that automation/AI really does have the power to begin significantly impacting upon workers income growth.

      After which the owners of the robot capital start scratching their heads and saying to themselves “gee….where did all the customers go?” Too bad robot “workers” can only produce while human workers fill the dual role of both producer and consumer.

      Automation may need to be managed to prevent it from suppressing growth and creating recessionary conditions.

    24. Damien Van Brink says:

      Interesting analysis Bill,

      Most boffins are predicting the short term loss of jobs to AI, barely any are making the bold (and seemingly incorrect) claim that mass job loss to AI has been happening. Your analysis looks at the latter claim, which we both agree is incorrect.

      The hype is hype, but the claims are sensible. Predictions of mass replacement are a result of projecting the rate of improvement in robotics and deep learning systems into the future. Only the i incredibly obstinate or unimaginative deny that self driving cars/trucks will be a huge disruptive force in the next 10 years. The technology is essentially finessed already, sans a few anomalous traffic situations.

      The hyperbole enters around discussions of mass AI incursion in the short to medium term. This may turn out to be true, but deep learning will need a revolution in capability, where as future projections rely on the pace of evolution.

      Of course, even the current unemployment levels are unacceptable (or the welfare response is anyway), so we don’t need to wait for the robots to come before implementing a JG or UBI.

      FYI I was in the Q&A audience with Bill Shorten as solitary guest and was shortlisted to ask a question about JG and UBI, I was bumped to position 19 and never got to ask the questions. Instead we got peppered with vacuous questions like “when will we see the real Malcom”.

    25. Some Guy says:

      Jim Green: There is no reason why a JG cannot go hand in hand with UBI.

      Yes, there is, because a UBI can’t go hand in hand with anything, or itself. A UBI that actually does what the children’s crusade behind it believes is just not possible, ever, anywhere. You wanna see a failing and punitive welfare system? Get a UBI. :-)

    26. Kevin Harding says:

      Again disappointing level of critique about UBI,willfully so it seems .
      The devil is in the detail and by far the most important detail as MMT iers
      should understand is the fiscal context.Any progressive should support UBI
      if it is a part of fiscal stimulus and fiscal transfer strategy.
      It is possible to devise a UBI which will make the poor poorer but it is also easy to
      devise a JG which provides cheap labour to exploitative employers.Slavery
      is the most common historical JG.
      Both can be a part of a progressive full employment strategy if they raise aggregate demand
      and direct more money to the working poor.

    27. sean says:


      Most research that comes into contact with complex systems ends up on the “future work treadmill,” where the exciting breakthroughs are always over the next horizon. Or to be more blunt, AI is always fifteen years away.

      Now it’s difficulty to compress an eight-year computer science curriculum into a comment on a blog post, and even then most other CS PhDs I know are all hype about deep learning, but… We learn in our theory of computation courses that it only takes three operations to make a computer: the not-and Boolean operator, write to memory, and read from memory. Not-and is a linear operator; it can be written solely in terms of matrix multiplication. Because of this, the only problem computers solve is linear algebra. (Square roots and the like are approximated with series expansion.)

      The implicit claim in the AI hype, and the autonomous car hype and even in numerical simulation, is that intelligence and driving and the ocean and the stars are all (high-dimensional) linear systems. Because this obviously isn’t the case, we’ll be stuck on the future work treadmill for the foreseeable future.

      There are deeper problems relating to the limitations of mechanical process—it turns out that it’s impossible to do logic by rote—but the thing about linear systems is probably the most accessible. Or y’know, I might just not have the imagination to overcome my education.

    28. Leftwinghillbillyprospector says:

      @sean – I still feel that the threats are real to a certain extent. Based on what I’ve heard at least.

      Is it true that algorithms are now sufficiently advanced that no human being – not even the worlds grandmasters – can possibly match them in strategy games like chess and go? (and is my terminology correct here?)

      It’s easy to understand how people could feel intimidated. And as I said earlier, I see potential for capital to use this as a powerful industrial weapon.

    29. Damien Van Brink says:

      @ Sean

      AI is not 15 years away, is already here, in our search engines, our navigation systems and image recognition algorithms. Are you meaning to use the term general AI perhaps? If you use a poorly defined term like AI then you can always make broad statements about it’s applications and development.
      I made the point that people are projecting future applications based on EVOLUTION, not REVOLUTIONS (or “breakthroughs” as you put it). We don’t need a revolution for significant job loss to occur, although obviously that would increase the magnitude.

      I appreciate your knowledge in the area however please excuse me when I largely discount it. I have an honour in physics (far from a PHD i’ll admit) however I would never pretend to be an expert in cutting edge geophysics or fluid dynamics. Unless your thesis was on deep learning then I would suggest increased humility. Not trying to be mean, but I have noticed that CS and MBA grads are among the most arrogant, physics/bio students have a stronger concept of epistemology as they realise how much the dont know.

      Looking forward to your further constructive disagreement.

    30. BobbyT says:


      You claim “The implicit claim in the AI hype…that intelligence…are all (high-dimensional) linear systems. Because this obviously isn’t the case”

      Bzz! Wrong. Quantum states evolve linearly and in general operators in QM are linear. Reality at its most fundamental level is consistent with a linear physical theory. The apparent non-linearities are just that—apparent—and disappear when you zoom in close enough.

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    To prove you're a person (not a spam script), type the answer to the math equation shown in the picture.
    Anti-spam equation