Robotics Replacing Soldiers

Pilots and drivers are clearly on the wane in modern armies. How long will human infantry last? The Pentagon is pouring resources into robotic advancement. Like the way many industrial enterprises are going, the wars of the future will involve a network of integrated robots supported by a few highly-specialized human operators and technicians.

Check out the exoskeleton concept (YouTube):
'Ironman' Exoskeleton
High Power Robotic Exoskeleton
For a while, the technology doesn't threaten the soldier's position, and will enable him to do more. But technology that can leverage a human's capabilities usually ends up replacing the human worker. At the end of the second video we see prospective models for humanoid battle suits that could become autonomous someday: 'if you step out of it, it becomes a humanoid robot' we're told.

For the next decade or so, human infantry will be training the robots by machine learning how to fight and move in battlefield environments just by inhabiting them. Sooner or later, there will be a benefit to ejecting as many wobbly, vulnerable human internals as possible and allowing the robots to do the rampaging by themselves.

The future robots of war will have more accuracy, reaction speed and stamina than humans are capable of. 
They will also have less fear, indecision and remorse. Mobility is currently holding back their application,  which is why there is such a drive at the moment to overcome this problem. What they lack in general mobility, they currently make up for with incredible niche capabilities -- some robots now are good for leaping 30 foot fences, others can navigate small tubes. They are immune to radiation sickness, smoke inhalation and bio-toxins.

Warfare is not an industry I support anywhere in the world. Nevertheless, it is worth noting that redundancy of human infantry will have a sharp impact on inequality. Recruits are usually poor and unskilled, and a career in the army is often one of the few means for upward mobility for many families. A dramatic reduction of infantry soldiers would disproportionately impact blacks and Latinos in the U.S., demographics that are already hurting badly with the decimation in manufacturing jobs.

Also, these videos are interesting. They give you an idea of the breadth of recent experimentation going on unmanned surveillance and killing machines:
 
Two Decades of Unmanned Ground Vehicles Compilation

Unmannned Air Vehicles Compilation

Andrew McAfee: Race Against the Machine

Andrew McAfee, coauthor of Race Against the Machine did very interesting TED talk recently on the subject of his book, overlapping heavily with the subject of this blog. The first half of the talk does an excellent job of stating the problem. He demonstrates how businesses are profiting more than ever today, investing in technology more than ever -- but are not hiring. He also talks about how various technologies are starting to catch up and even eclipse humans in cognitive tasks such as translation and writing grammatically-perfect articles, and that 'we ain't seen nothing yet'.


Oddly, while demonstrating that 'droids are coming for your jobs' he also claims that current debate on 'whether these technologies are affecting people's ability to make a living' is 'missing the point entirely'.

For the millions of people people losing their houses to foreclosure, shuffling in the dole queues and soup kitchens, collecting paltry benefits, wallowing in depression and freezing in alleys and tent cities I would say the concern about losing income is exactly on point and in need of urgent focus. Is he saying that we ought to just wait out the storms in a 'laissez-faire' kind of manner and that the problems will spontaneously sort themselves out?

Unfortunately, I think so. He doesn't offer anything in the way of suggested actions and decisions we could be making. Instead, his talk pivots half-way through onto an optimistic feel-good frame which I think has shaky foundations. He points to one good aspect of technology at the bottom of the pyramid -- citing a study of poor rural fishing villages leveraging the power of cell phones to improve their knowledge of the market situation and reduce waste.

Yes, we know that human productivity rises when technology is introduced, and it's initially a huge benefit to independent workers. But what happens when poor fishermen with creaky boats come into competition with hyper-efficient automated fishing operations? On land, where robots have a surer footing, we can see how small, poor farmers suffer in competition with gigantic, scaled and highly automated agribusiness.

A time when hands are not needed to steer the ship and haul the fishing nets can't be too far away, as these are pretty routine tasks, and there's plenty of research going into ship automation. With no income stream to leverage,  information carrier technology becomes increasingly useless, not to mention unaffordable (in the current consumer paradigm).

I'm too am hopeful that technology will benefit the neediest in the world. We know of it's potential, but we can't afford to think that technology will steer itself to that end without heavy political upheaval going on of some kind or another. That's the missing step in the transition that too few people (and almost no prominent futurists and economists) are willing to talk about -- including myself, for now.

U.S. Unemployment is in Fact Getting Worse

Obama and much of the US media has recently managed to paint a bright picture of jobs recovery this quarter. Unfortunately, reality begs to differ. The perception was manufactured with clever manipulation of the data. This fantastic analysis, by Daniel R. Amerman at Financial Sense takes a candle to the lie. From the Article:

"When we pierce through this statistical smoke and mirrors and factor back in [...] 9 million jobless whom the government has defined out of existence, then the true unemployment rate is 19.9% and rising, and not 8.3% and falling. [...]

Since 2007, the Constant Workforce Participation unemployment rate has risen from 9.6% to 19.9%. This increase of 10.3% for a labor pool of 160 million people means 16.5 million additional jobless people since 2007.


The difference between 6.7 million missing jobs with a rapidly improving jobs picture, and 16.5 million missing jobs with unemployment still rising, is the difference between night and day. It is the difference between a recession being brought under control - and a depression that continues to worsen."

Obsopocalyptic Visions

Miles of sophisticated capital devoid of human labour... Automated factories, automated stacking for shipping, automated warehousing...

With automated haulage (possible with the wave of unmanned vehicle technology in recent years) and automated retail on the horizon, we're getting close to the point where you can have a gigantic distribution system churning out products with nary a human involved. Unless you count the owner.

With such awesome capabilities, is it any wonder that profits are skyrocketing for capital, while unemployment remains intractable, and may even be getting worse?

Automated Manufacturing

Automated Palletizing

Automated Warehousing

Can We Really Depend on Hi-Tech Start-Ups to Power Future Employment?

Many who refute the coming obsopocalypse argue that displacement by technology simply frees people up to do new jobs in more cutting-edge ventures -- ventures often created by the very same disruptive technology. Obviously there new businesses and new jobs categories pop up all the time -- but the related jobs are just not being created in large enough numbers any more.
 
The Kauffman Foundation produced an excellent study last year Starting Smaller; Staying Smaller, America's Slow Leak in Job Creation detailing how new start-up businesses have become increasingly poor engines of employment.

From the report:

''Even before the Great Recession, firms were starting smaller. They were opening their doors with fewer workers than the historic norm and were relatively reluctant to expand their workforces even during good economic times. Since at least the middle of the last decade and perhaps earlier, the growth trajectories and survival rates for these businesses meant that they were contributing fewer and fewer new jobs to the  economy.
[...]
Media and academic commentators who bemoan America’s unusually slow rate of job creation after the 2007–2009 recession are missing what we believe is a longer-term trend that began earlier in the decade and might best be called a slow jobs “leak.”
[...]
In many cases, companies or individuals that once would have been hired as employees of a business now are performing the work on a temporary basis as contractors through other professional service organizations or under their own self-employment contracts. These individuals, while sometimes characterized as "entrepreneurs,” are not likely to employ others or to reach significant scale".

Some of key trends from the report:
-The number of new start-ups has declined significantly in recent years
-The percentage of the new business that are providing jobs for other people (other than the founder) is decreasing -- there's a huge rise of one-person businesses.
-The average number of employees per new business has been declining since 2002 (when 10.8 would be employed) to less than 8 now
-Fewer new businesses are surviving more than 5 years
-Of  the businesses that survive their initial years, and go on to grow their operations, job growth is now slower than the historic norm

The extinct switchboard operator

The report refrained from speculating on the causes of the job growth leaks compared to previous decades. To me, and I believe to anyone who gives it 5 minutes thought, the story seems quite obvious. The timeline completely coincides with the computer revolution, and the growing use of computers makes an increasing number of former roles redundant.

The precipitous decline in job creation of contemporary start-ups is an entirely predictable result of technological development. Advancing technology, especially IT tools in the past 2 decades, increases worker productivity, allowing companies to do more with less workers.

For example, secretaries, data entry staff and even accountants are found less and less in new companies. The notion of a human secretary at a desk could soon seem as quaint as telephone switchboard operators today.

On Software Automation (and Dilbert)

Serious point to these comics -- automation used to be mainly associated with manual tasks, but we've seen a steady march into elements of office service work that secretaries, call centre operators, bank tellers etc. used to do. In many companies now, even CVs and job applications are being read and discarded automatically by clever software identifying the best fits, so that only a shortlist, a fraction, of the applications is passed to a human for the final decision.

Pretty much any clerical task that can be done using a computer or network can be automated by sufficiently advanced software. Because software is replicable, it is stunningly cheap compared to hiring human brains. Unlike machinery and people, software does not really wear out over time, and it is can be very quickly upgraded. Institutions in the service sector are becoming increasingly lean with IT technologies, a trend that started in the early 90. This is probably why the disruptive start-ups of the internet age don't create as many jobs as they destroy.

Anyway, I've always loved Dilbert, enjoy :)




 



(From Scott Adam's Dilbert cartoon,  www.dilbert.com)

The Messenger: Martin Ford

I doubt that the threat of job displacement by automation is far from the minds of millions of workers globally. But if you were to base your opinion solely on what the majority of economists, captains of technology, and futurists are saying you would probably miss the story completely. Technological unemployment does not loom large on the mainstream radar, just as wealth inequality was a non-issue until popular unrest exploded onto the streets in recent years.

One exception to the prevailing  myopia is Martin Ford whose timely book Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future  is probably the best extended analysis on the issue to come out this decade.

As a business owner of a tech company, educated and experienced in software trends and economics, Martin Ford is well placed to properly observe the effect of technology on the future economy. More so, perhaps, than the specialized economists and futurists who disagree with him. His breadth of understanding in both fields is evident from the manner in which he anticipates arguments from all quarters.  

Ford has clearly taken pains to make his book accessible to the public in an effort to raise consciousness on the issue. Without oversimplifying the issues, the book is written with enough clarity that even those unfamiliar with economic and technological trends are sure to find this an absorbing read.

Since his tightly-written arguments are difficult to summarize without loss, I'd rather his work speak for itself. Here are a few key snippets (slightly out of order) from Lights in the Tunnel which lay out the core of the problem. He has a lot more to say, so do read the whole thing if you get a chance.

From Lights in the Tunnel:

On The Growing Reach of Automation

‘The continuing advance of computer technology along a geometrically increasing path and the diminishing returns from investment in education seem to make a very strong case that the average worker—and perhaps many above-average workers—are in clear danger of having their jobs automated’. (p 54)

‘The conventional wisdom as generally presented by economists and other analysts is that technology creates jobs. While history has shown that this is indeed true, it also shows quite clearly that the new job types created by technology are very often themselves quickly vaporized by the same phenomenon. The IT jobs that are now being offshored and automated are brand new jobs that were largely created in the tech boom of the 1990s’. (p 57)

‘The fact is that the vast majority of our workers continue to be employed in traditional jobs. The new job types created by technology represent a relatively small fraction of employment and […] often tend not to last very long. (p 61)

‘[E]ntire traditional job categories are at risk of being heavily automated in the not too distant future. To suggest that technology is going to somehow create completely new job categories capable of absorbing millions of workers displaced from traditional jobs is pure fantasy’. (p 62)

‘The specter of near fully automated supermarkets and chain retail stores is cause for genuine concern. These are now the jobs of last resort. These are the jobs that workers displaced from other industries take because there is nothing better available’. (p 79)

‘For knowledge workers, there is really a double dose of bad news. Not only are their jobs potentially easier to automate than other job types because no investment in mechanical equipment is required; but also, the financial incentive for getting rid of the job is significantly higher’ [because of the larger salaries]. (p 73)

On the Decline of Consumption and it’s Consequences

‘As a growing percentage of the population is exposed to direct evidence of ongoing job losses, many people will begin to experience a greatly heightened level of stress and worry. Facing this, individuals will take the obvious action: they will cut back on consumption, perhaps quite dramatically, and try to save more in anticipation of a very uncertain future’. (p 109)

‘In essence, we have succeeded in globalizing labor and capital, but we have really not globalized consumption. To a large extent, workers in low wage countries are not capable of purchasing the goods they are producing’. (p 113)

There is no incentive to produce products if there are no consumers with sufficient discretionary income to purchase those products. This is true even if intelligent machines someday become super-efficient producers. […] If we consider the singularity in this context, then is it really something that will necessarily push us forward exponentially? Or could it in actuality lead to rapid economic decline? (p112)