Please forgive me. This is a very long posting. It took a long time to write, so reading it will be less of a commitment than I’ve put into it. Why so long? Well, folks didn’t believe “the short form”, so I’ve had to elaborate and provide a theoretical underpinning. That takes more words than “does not” ;-)
Several folks have expressed great angst over robots and manufacturing. Generally expressing the position that robots are dooming employment (often with a subtext implied of ‘in a capitalist market’ as opposed to the ‘make work’ world of governments) and that it is inevitable that the economy will collapse as the excess workers pile up without any paycheck with which to buy the goods and services produced.
Others, me among them, point to a very long history of mechanization of labor and how the economy has always expanded and living standards risen after more automation and mechanization are implemented. This history starts far far before robotics and computers; though they are oddly connected. It is an interesting thread that starts with, well, thread. But we’ll come back to that later.
The more obvious examples I’ve posted before (usually in comments). Farms at one time employed 90+% of labor in the USA. Now down in low single digit percents. My Grandfather was one of those farmers. His son, my Dad, did not become a farmer. Yet before being a farmer, Granddad was a blacksmith. His father (my Great Grandfather) and several before him back into the 1700s at least (that we know for certain) were all blacksmiths. The furthest back we document is to the 1700s where a guy got off a boat in Virgina and worked iron, so was a Smith. Thus my last name.
A Personal Historical Perspective
How did a smith end up a farmer? Mechanization of ironworking. The industrial revolution. It is important to realize that a factory is nothing more than a giant robot for the manufacture of goods. Historical factories did not have computers running the machines, but the “program” was embodied in the structure of the factory. A giant machine with the flow of raw materials in, the processing, and the formation of finished goods at the other end all programmed by the design and structure of this giant machine for fabrication. It absolutely decimated “craft production” in all sorts of industries. Smithing among them.
There were great wailings and gnashing of teeth over all the people being put out of work and how horrible it would be and how the economy would never recover. It was traumatic. I’m sure several generations of smith were not expecting that THE premier craft of the day would become obsolete. The Smith was not just the maker of horseshoes, but the locksmith and war industry as well. Often orders were given to “capture the smith’s alive” as that was capturing the weapons manufacturing industrial base. It was that important a job. Yet a machine forge that does nothing but make nails and screws and screwdrivers all day long does it incredibly more efficiently. An hour of high skill labor to make a good screwdriver is, today, a $1 item at the hardware store – and it is a better product too. So the smiths “moved on” and Granddad expanded the farming operation. But he did keep a small smithy on the farm and did smithing for the local Amish. That’s where Grandma came from. ;-)
It wasn’t just smiths. Chandlers and Coopers and Tinkers and Wrights – all the huge laundry list of last names that were once connected with craft manufacture were all put out of a job. Look around you at your family, friends, neighbors, and all. How many of them have such manufacturing linked names? All their families were out of work. Yet the world did not end. The economy did not collapse. Life went on. We got our buggy whips made by machines instead of by hand and our candles uniform and much cheaper instead of laboriously hand dipped by the Chandlers down the street. (The Amish side of the family really appreciated the cheaper candles). While Granddad mostly made a living out of farming.
Then mechanization came to the farm. The tending of large draft horses gave way to the tractor. Usually unmentioned is all the hundreds of hours and dozens of acres it takes to tend, feed, and propagate draft horses. The Amish still do it, and it can be an effective way to farm, but it chews up resources at an astounding rate compared to a tractor and some Diesel fuel. “Biomass” is not a new fuel. It has been run through horses and oxen for thousands of years. It takes a enormous amount of land and water and fertilizer and labor to keep that team fed. It is the failure of “biomass” as fuel compared to Diesel that caused tractors to dominate. Now one farmer could farm far more acres, and you didn’t have to set 20% or 30% aside for fuel for the team. Ag output started to rise dramatically, on ever less labor. So Dad didn’t stay on the farm. He “moved on”.
Where Dad went isn’t quite as important. He moved into various ‘services’ mostly. First as soldier in W.W.II and then a brief stint building houses in the booming California. But most of his years post war were spent running restaurants or selling and managing realestate. Manufacturing, farming, and smithing left far behind. He did plant a very large garden and assure we kids knew how to farm and garden (somewhat different, BTW) and had me help him dig a well in the back yard so I’d know how to do it all (that Amish thing). He also taught me basic smithing as he felt it important that the craft not die out. Just as his Dad had taught him. With many generations prior the same. But he made zero money from farming or from smithing. (We did eat well from the garden and from the 5 acres outside town where he had his cows and ‘toy farm’ that kept the freezer full of beef).
On Mom’s side, it was mostly sailors. “Several” generations. How many is not known. At least 3 generations of Joseph Sumner who’s log books we know of. Then Mom married a Yank during that W.W.II dust up and came to America. Back in England, the rest of the family (including all the boys) did not go into sailing. Why? Mechanized ships did not need the huge crews of a 3 masted sailing ship. Nor did they need the riggers, the carpenters, the hoard of folks needed to build and repair wooden ships. Massive layoffs due to the “robot” of a mechanical ship that did things with the turn of a wheel and movement of a lever instead of a mass of men climbing in rigging and shifting spars. Instead, many of those folks became skilled in jobs like riveting and painting and electrical work. My Uncle one of them. He went into that newfangled radio stuff. (Over on this side, some of Dad’s brothers went into domestic electrical work and have a large Electrician Contractor company in Iowa).
The point behind all that family narrative?
Things change. People change too. Careers are no longer just learned from Dad and that’s that. What you do for a living now changes as the technology changes, and has since at least 1600. I’ve had several different jobs in my life. From stacking cases of peaches in a cannery, production planner for semiconductors (that didn’t exist when I was born), to various computer jobs. I expect I’ll have one or two more “career changes” before I’m done.
Which brings us to that thread.
The Jacquard Loom
A fellow wanted to automate even more the already advanced looms weaving fabric. The loom had greatly improved fabric quality and production with a mechanical powered loom that was far faster than by old hand methods. Yet it required some significant skill to get the different threads right to make a decent pattern. Jacquard wanted to change that and mechanize it.
Notice the dates in this quote. It was being designed in the late 1700’s and then demonstrated as a working machine in 1801.
The Jacquard loom is a mechanical loom, invented by Joseph Marie Jacquard, first demonstrated in 1801, that simplifies the process of manufacturing textiles with such complex patterns as brocade, damask and matelasse. The loom was controlled by a “chain of cards”, a number of punched cards, laced together into a continuous sequence. Multiple rows of holes were punched on each card, with one complete card corresponding to one row of the design. Several such paper cards, generally white in color, can be seen in the images below. Chains, like the much later paper tape, allowed sequences of any length to be constructed, not limited by the size of a card.
It is based on earlier inventions by the Frenchmen Basile Bouchon (1725), Jean Baptiste Falcon (1728) and Jacques Vaucanson (1740) A static display of a Jacquard loom is the centrepiece of the Musée des Tissus et des Arts décoratifs (fr) in Lyon. Live displays of a Jacquard loom are available at a few private museums around Lyon and also twice a day at La Maison des Canuts, as well as at other locations around the world.
To those who have “run ahead in the book”, yes, those “punched cards” were taken up as a control mechanism by “computers” in the early days and on up into the 1970s when I learned to operate an 029 Keypunch in college. (BTW, there were thousands employed then as Keypunch Operators, now largely gone on to other work. Though you can still see them in classic movies from the ’60s era, one including Doris Day.) The history of computers and automation reach back to the 1700s and 1800s. We’ve been at this process for at least 200 years now. You would think that if it were going to cause catastrophic and unrecoverable collapse of the economy it would have done so by now.
In comments here: https://chiefio.wordpress.com/2015/10/12/the-fu-factor/#comment-64842
Queen Elizabeth I arguably delayed the “Industrial Revolution” by 140 years because she wanted to avoid killing the jobs of the knitters who made hosiery in the 16th century.
This site has an interesting history of knitting. I care about knitting partly as my Mum taught me to knit at about 4 years old. I was curious about this thing she was always doing, and she decided to give-in and teach me this “girls work”. I still like knitting, but have not done it in a long while. Mum had been a seamstress in W.W.II making uniforms. Later she was a waitress in our restaurant. She also had several other jobs including census taker. So she, too, had transitioned from “manufacturing” to other work.
A primitive hand knitting technique dating back into prehistoric times was reproduced using one needle and short pieces of fiber in a study by Dorothy Burnham in the Royal Ontario Museum in early 1970’s. The next documentation speaks of three pieces of hand knitting found at the site of the old Syrian City of Dura Europos, which was sacked by the Persians in 256 AD. Study of these items defined them as having been hand knit with two knitting needles. Evolution has been and is ongoing from handknitting with one to two needles and onward to peg frames, knitting frames, circular sock machines, and home knitting machines. Just to spur your interest, I will give you a few tidbits of information from each stage.
Note that in the original it said “had knitting technique” and I added the ‘fix’ of the obvious dropped n.
Knitting has been around since “pre-history” with actual cloth samples from 256 AD. Then note the list of automations. Peg boards. Frames. “Circular sock machines”. Does that last one make you think of 1970s and South Carolina? Or 1990 and China?
The Crane Knitter ~ Flat Bed and Circular Sock Machines ~ 1870
The Reference in Cassell’s Household Guide of 1870 advertised the Crane Knitter as a striking new American invention: “It had a massive metal plate keyboard in which deep, square grooves for the reception of 106 needles are cut, and “asort” of metal carriage or traveler, which can be moved forward or back on the keyboard with the greatest facility by the aid of a short ball-crowned handle.” This was one of several sock knitting machines of that era, which included both flat and circular sock machines.
I am proud to say that I have two circular sock machines at The Knit Tree which have been restored. Occasionally, I have the opportunity to knit a special pair of socks from these sock machines for the One of a Kind category of our online catalog. Go see if a pair is there, now.
My Anslie and Legare are pictured at the beginning of this article (c late 1800’s) When I found the Anslie machine, (in the attic of one of my students) it still had a “Work Sock” hanging on it. The cylinder was jammed and it was inoperable. With a loving restoration, this sock knitter was brought back to life. A box of labels “Bundles for Britain” and the hand written pattern used for these socks which were knit for our soldiers during the war, were also in the wooden carrying box. This is the same instruction pattern, I follow for the socks offered in The Knit Tree’s One of a Kind Showroom – making these socks a special treat!
Yes, 1870. We’ve had mechanical knitters for the home since at least then. BTW, that “Bundles for Britain” may have helped keep my spouse’s Dad and his squad alive. When during W.W.II, they were wandering around Europe as the 101st Airborn, frozen or molding feet was still a major cause of casualties. He managed to “cut a deal” somehow ( Sargents have their ways ;-) and managed to get one big box of socks from Britain for his crew. He had been the British Liaison Officer to the Gliders and after landing may have used that to some advantage… but I digress.
Curiously, there is another connection of computers to thread here. Knitting Patterns are written in an arcane symbol method. It includes processes, particular steps to perform, and groups of repeated actions gathered together in one chunk, and repeated a number of times. In all reasonable ways, this is a programming language. Complete with subroutines and function calls. I used knitting book patterns as an example of writing a program in my teaching of Introduction To Computers. That, and that Ada Lovelace was the first programmer tends to get the women in the class more interested. Women have been doing programming far longer than men.
But earlier, and to GallopingCamel’s point:
Stocking Loom ~ 1589
Knitted stockings became high fashion in the 16th century. Wishing to please Elizabeth I, Rev William Lee, of Calveron, Nottingham, designed his stocking frame to produce fine silk stockings. This Stocking Frame did not impress the Queen. It is said that she was much to distracted with their Armada victory to take notice of Rev Lee’s invention. Hand knitters opposed the idea of knitting frames, therefore, framework knitting evolved very quietly for the next century.
If Elizabeth I had been a bit more enthusiastic, Britain might well have had a knitting revolution in the late 1500’s and the Jaquard Loom might have a more British name attached.
We see this consistently throughout economic history. Those who advance mechanization / automation / roboticizing (whatever you call the technological advance) end up rich and with lots of well paying jobs. Those who eschew it end up unemployed wondering what happened. It is essential to “embrace change”. As I’ve mentioned before, especially so in Silicon Valley where whole industries spring up, grow old, and die in a decade or two. Yet The Valley has more nearly full employment than just about anywhere else and with higher paying jobs. That is the result, not the cause. I could just as easily have ended up a cannery worker instead of a computer programmer; but the cannery moved away and automated their operations in a low labor cost land.
That site is mostly dedicated to home knitting machines. Now, thanks to other automation, you can buy them for “toy prices”. When made by hand by the smiths of the 1700s, they were incredibly expensive machines that only the rich could afford. This one, from Amazon, costs under $30.
NKOK Singer Knitting Machine
List Price: $31.99
Price: $28.67 & FREE Shipping on orders over $35. Details
You Save: $3.32 (10%)
Only 4 left in stock.
Sold by Country Supplier and Fulfilled by Amazon. Gift-wrap available.
Want it Saturday, Oct. 24? Order within 13 hrs 49 mins and choose Saturday Delivery at checkout. Details
Nkok-Singer knitting machine
Perfect for beginner knitters! knit cool fashion accessories like hats scarves socks and leg warmers!
Recommended for ages 8 and up
Want a knit cap or socks? Now it’s nearly trivial to do. Or you can buy them made on large machines in China for low prices. We all benefit when socks are no longer something you need to “cut a deal” to get and are instead so common place we no longer even think about them.
FWIW, I have a pair of hand knit socks that I treasure. My Mum made them for me in the ’70s for when I was going skiing. They are warm and bulky and work well in large boots. They are a treasure to me, rarely worn anymore. Yet the simple fact is that machine knit socks are better and stay up without frequent tugging… and it didn’t take a few weeks to make a pair. So I have one pair of hand knit socks, and have gone through hundreds of machine knit in the same time. Nobody even knows why you would want to ‘darn a sock’ anymore. That’s a feature.
You can write substantially the same story about every industrial manufacture over the last 500 years. A very long history of mechanization, automation, and increasing prosperity with ever more “stuff” for everyone at ever lower prices. Yes, we’ve had to change jobs. The Chandlers who live down the street are more likely to sell insurance or repair cars than spend all day dipping candles. I know of at least one clan of Smiths who do computer programming… The Coopers usually don’t make barrels any more, but Cooper Tires is a major maker of tires (on automated machines). We have moved, ever more strongly, into being ‘managers of machines’ instead of brute labor. Even my mechanic has a tire mounting machine used to remove old tires, mount new ones, and with a computerized automated wheel balancer that tells him just where to tap on the wheel weight and how big. I remember watching the local mechanic do that all by hand when I was a kid. A very long and hard process. Now just a minute or two per wheel. He benefits by more total billings in a day and I benefit by a better balanced tire and lower prices.
About Amdahl and Gustafson
Moving away from several hundred years of “existence proof” in the history of automation, is there some theoretical basis for not worrying so much about robotic labor?
I think there is, but I doubt it has been stated this way before.
First, a digression into Amdahl’s Law. In full disclosure, I worked at Amdahl Corp for a few years in the 1980’s. They made giant mainframe computers that cost $Millions each. These were of a kind called “scalar processors” in that they do each job in a linear series of steps. This is what most CPUs (Central Processing Units) do in most computers.
There was another kind of processor taking off then. Cray was making “vector processors”. These would do a whole bunch of the same operation on a batch of data all at once. So you would fill an array (data structure) with, say, 64 values for hours worked, and another with 64 values of “pay rate” and then ONE instruction would do all 64 multiplications at the same time and give ONE big array of 64 results. (That 64, btw, was the width of the vector unit in “my Cray” when I ran one. It was called a ‘stride’, and lots of Cray FORTRAN was written with array arithmetic and loops of “one stride” or 64 data items as a chunk. This let the compiler optimize for the vector unit) The “vector unit” survives as the GPU or Graphics Processor Unit in most computers with fancy screens and graphics, like your PC or Macintosh. Early PCs had no GPU. Companies like Nvidia changed that and made millions in the process.
At the time, there were intense arguments that were asserting the Mainframe was not relevant as parallel processing would consume their whole market. A hoard of smaller computers and / or vector units would make a Big Iron machine irrelevant. Amdahl showed why that was not the case. To this day you can still buy “Mainframe” computers, even though they are many 10s of times more powerful than the ones from the 1980. We still need “big iron”. But why?
Amdahl’s law, also known as Amdahl’s argument, is used to find the maximum expected improvement to an overall system when only part of the system is improved. It is often used in parallel computing to predict the theoretical maximum speedup using multiple processors. The law is named after computer architect Gene Amdahl, and was presented at the AFIPS Spring Joint Computer Conference in 1967.
The speedup of a program using multiple processors in parallel computing is limited by the time needed for the sequential fraction of the program. For example, if a program needs 20 hours using a single processor core, and a particular portion of the program which takes one hour to execute cannot be parallelized, while the remaining 19 hours (95%) of execution time can be parallelized, then regardless of how many processors are devoted to a parallelized execution of this program, the minimum execution time cannot be less than that critical one hour. Hence, the theoretical speedup is limited to at most 20×.
It is my assertion (we can call it “Smith’s Corollary to Amdahl’s Law”) is that something very similar applies to the economics of automation. I would state it as:
“If you automated all that can be automated, the work that cannot be automated still remains, along with the work of tending the machines. This limits the total work displacement from automation.”
Essentially, take that “improvement to an overall system” and apply it where the “system” is the economic performance and work done and the “improvement” is the robots. Once they’ve been applied to all they can do, the rest dominates the “work space”.
Here’s the graph from the Amdahl’s law link:
I would assert the same holds for automation. IFF 95% of some manufacture can be automated, we will use a couple of thousand machines to do it, and that will make 20x as much stuff. We tend not to make the same stuff in the same quantity, so just as many folks will be employed in the non-automated parts of that particular product (shipping, accounting, marketing, sales, machine maintenance, computer programming, etc.)
Now couple that with the REASON you automate being lower costs to produce, and the “law of supply and demand” then those lower cost goods will be in ever more demand, leading to much more production, leading to expansion of the work needed in the non-automated portion. Someone is repairing, installing, and servicing all those knitting machines making all those stockings world wide. Think back to W.W.II era. “Silk Stockings” cost a minor fortune. Then nylon was invented, and matched with knitting machines, and generations of guys and gals have enjoyed the results ever since. Even in the ’60s I remember a ‘run in a stocking’ was cause for pangs as they were a costly item. Now they are nearly disposable. Yet also nearly indestructible.
In the end, instead of needing a few thousand workers making silk stockings (from tending moths to knitting a million knits / stocking) as they cost so much to make, we have a largely automated process that still employees lots of people that I’d guess was up in the millions if you include sales channels.
Substantially the same thing can be said about nails, hammers, hand tools in general, dishes, buckets and so much more. I don’t spend half a day making a hammer. I buy one for a fraction of an hours pay. If it rusts I don’t spend 1/4 of a day re-making it and cleaning it and all. It goes in the recycle and I buy another one. Thus the non-automated part of that hammer manufacture ends up consuming just as much labor as ever as we have more hammers (by a long shot) than when they were hand made.
This has shown up in food production in an indirect way. We eat more beef. Back in the 1500s and even into the 1800s, people ate a lot more grain and a lot less meat. Now we have mechanized agriculture. This has made grain so plentiful that we feed it to animals. Cows have about a 10 : 1 “feed conversion”, so you feed them 10 lbs of dry grain to get one ‘wet pound’ of meat. That 10 lbs of corn would feed a person for 10 days. That one pound steak gets them through one very nice dinner. Chickens are about a 2 or 3 : 1 conversion, so “poor people” now eat a lot of chicken. Long gone are the days when “A chicken in every pot” was a promise of great wealth to come. When I was about 6, it was still common to think of a chicken as a ‘special Sunday meal’ instead of ‘oh that cheap stuff again’. Automated chicken farms and automated chicken plucking and butchering changed all that. So we eat a lot more chicken and let them eat the grain. Total people employed in producing chicken has gone up.
BTW, I’ve plucked a pheasant by hand. If there is any job in the world I think is “not worth having”, it is chicken plucker. Lord thank you for the automated plucking machine!
On the one hand, Amdahl’s Law and Smith’s Corollary predict that as automation proceeds, there is a limit to what can be automated and the non-automated part comes to dominate the work. As we are already about 200 to 400 years into this whole automation thing, I’d assert we are already well along that graph.
Then add in lower cost to manufacture and how that interacts with supply and demand, that means more total consumption, and so more total production. This “economic growth” leads to more employment in those fields than the “static scored” graph would predict. And that, too, is what has been seen in the history of the world.
From knit shirts and stockings, to cars, radios, and even chickens: The more automation, the more of it we buy. Even computers themselves. In 1960 nobody had a personal computer. Now I’ve got 5 of them “hooked up” and 4 more sitting around archived as “obsolete”. And that doesn’t count my tablet and cell phone. In 1960 we had ONE wired ‘black phone’ for the house. Now we’ve got an internet connection AND one cell phone per person AND I’ve got 2 portable wireless hotspots. Communications is now very automated and very computerized, and I’m consuming far more it than ever before.
So not only does the non-automated part come to dominate the industry for any given industry, but those industries that automate most, grow most. Leading to more consumption of goods and more employment.
Flys In Amdahl’s Soup?
There’s another “law” that I think is also important. Gustafson’s.
But I think it is important for the context of what it says more than for the actual words.
Gustafson’s Law (also known as Gustafson–Barsis’ law) is a law in computer science which says that computations involving arbitrarily large data sets can be efficiently parallelized. Gustafson’s Law provides a counterpoint to Amdahl’s law, which describes a limit on the speed-up that parallelization can provide, given a fixed data set size. Gustafson’s law was first described by John L. Gustafson and his colleague Edwin H. Barsis:
It then goes into a bit of math that those inclined can go look up. The basic notion behind Gustafson’s Law is that the problem changes. It isn’t a fixed pie of problem size that the computer is attacking. So make a massively parallel computer, folks will find some massively parallel problems to use it on.
That, too, is exactly what has happened.
These two are often presented as fighting each other. I think that is not so.
Amdahl is talking about any one problem being worked and Gustafson is talking about all potential problems that could be worked.
Two different problem spaces.
Gustafson’s law addresses the shortcomings of Amdahl’s law, which does not fully exploit the computing power that becomes available as the number of machines increases. Gustafson’s Law instead proposes that programmers tend to set the size of problems to use the available equipment to solve problems within a practical fixed time. Therefore, if faster (more parallel) equipment is available, larger problems can be solved in the same time.
Accordingly, Gustafson called his metric scaled speedup, because in the above expression S(P) is the ratio of the total, single-process execution time to the per-process parallel execution time; the former scales with P, while the latter is assumed fixed or nearly so. This is in contrast to Amdahl’s Law, which takes the single-process execution time to be the fixed quantity, and compares it to a shrinking per-process parallel execution time. Thus, Amdahl’s law is based on the assumption of a fixed problem size: it assumes the overall workload of a program does not change with respect to machine size (i.e., the number of processors). Both laws assume the parallelizable part is evenly distributed over P processors.
Applied to the economics of automation and robotics, this is, to my point of view, basically the part about “we do more of it and buy more”. As we apply robotic assembly to things, we can make things we would never have made before. It is humanly impossible to construct an iPhone. Very large parts of it (or, in reality, very very small parts of it…) simply must be done by robots. Robotic assembly allows the iPhone to exist and so we buy lots of them. This employs folks all over the globe in the tens of thousands (and into the millions if you include the folks making aps, selling iPhones, iPhone network services, classes in programming IOS, and on and on). Yet we have millions of these exquisite machines that are in all ways possible a magic miracle that are thrown away when a new iteration comes out. Robots let that happen. The miracle becomes the mundane and we “do what could not be considered before”. Have disposable miracle boxes in our hands.
But it isn’t just computers and iPhones. Look at music. It was all hand crafted 500 years ago. Edison made it available to all, and music consumption went up. Robots run youtube. Yes, highly specialized robots we call “computers”, but they are robots. Disk controllers are indistinguishable from the controllers moving a grasping arm instead of a disk arm. Now we have $Billions in the “music” industry and every music recording needs a video with special effects, makeup, choreography, backup singers, etc. etc.
Similarly in movies. In the 1960s “special effects” were rare and expensive. Stunt doubles and various physical magic. The Foley Artist had a box of coconut shells for horse hooves, a cap gun, some bells and a wash board and other goods for making various sounds in the movies. Special effects and Foley Art were used lightly. Now they have a specialized automated sound robot that calls up any desired sound on demand, and even lets them make waveforms of any sort via computer synthesis. Now every movie has a list of tech support longer than the cast and you can’t find a movie without Foley Art used end to end and a load of computerized special effects all over it. In some cases, such as Avatar, most of the movie is ‘special effects’, or robotic production of the art. This used MORE people doing special effects, not less. Sure, a few carpenters likely had less set construction to do, but that didn’t cut budget of the movie spent on labor. There was a lot more electrical work on the sound stages.
As another example, my “old college roomie” is now teaching robotics at a local high school. They compete in robot competitions. This is not a small operation and they just placed at some level in the State competition. Not one of those jobs or the jobs related to all the events held would exist without the robots. The “problem space” has grown to include Robotics Engineer, Robotics Teacher, Robotics Judges, Robotics Venues, etc. etc.
Precisely because of Smith’s Corollary to Gustafson’s Law, that robotics and automation will lead to MORE uses in MORE ways and for MORE things than anticipated by static analysis of one use.
We see that playing out in the history of automation.
In the 1800s it was expensive and slow to cross the ocean as it took a Tall Ship with a large crew for along time. Steamers in the 1920s and 1940s made the journey faster with less crew, and more folks ‘made the crossing’. Instead of a ‘one way” when my Irish ancestors came to America, folks might cross a couple of times. Then, with the jet airplane the crew (labor) dropped a lot and the hours / passenger mile of crew costs dropped more. Now folks “jet over” for a long weekend. I’ve crossed the country (2800 miles) more times than I can remember. Once ever 2 weeks for a couple of years at one point. The fuel used to fly is far more than in a train. The difference is the highly automated cockpit and air traffic control and very low labor used. It is substantially an automated process compared to trains, and especially so compared to the trains of the 1800s. We can now have a flight crew of 2 pilots and a half dozen cabin crew take 1000 people coast to coast or from Europe to New York in about 5 hours. So we do it a whole lot more often than when travel was by manual operated ships, and trains that had similar crew sizes, but required a couple of weeks pay to cover them. It is a whole new “problem” that is being addressed. That of “weekend transcontinental travel”.
When you can look at nearly every industry and see the same thing, I think that’s a pretty strong existence proof for the fundamental nature of the economy. Not unlike Amdahl’s Law and Gustafson’s Law in computing, the advance of robotic production has the same effect on economic activity. Smith’s Corollary to Amdahl’s Law says that the non-robotic portion will come to dominate the labor space. (To some extent a tautology). More of us will be in “service” and sales and insurance and teaching and doing dance and putting on plays and polishing our speakers awards. Smith’s Corollary to Gustafson’s Law says we will be doing things we have no idea can be done today.
One final example. Especially on that last point. A relative just had surgery. They were out of the hospital in a day when I know that particular surgery takes a good week in the hospital (from my days working in hospitals on the wards). That surgery was done at a life phase where it would never have been done before due to it being “major surgery”. So what changed? It was done as robotic surgery via a very small incision. A great revolution in surgery is happening with robotic surgery. It is NOT reducing the number of surgeons. (Yes, you can quibble that it is robotic assisted surgery.. and I’d point out that iPhones are just robotic assisted assembly and socks are just robotic assisted manufacture and…) Since we don’t need holes big enough to get 4 hands inside a person, we can use robotic surgery and do what was impossible, at lower costs, higher speeds, faster recovery, better results. In the end, this means more surgery being done as the impossible becomes mundane.
It is that kind of thing that brings Smith’s Corollary to Gustafson’s Law into force. As long as there are unsolved problems to solve, there is no risk of running out of work for computers (scalar or vector) and as long as there are unmet human desires there is no risk of running out of work for robots and people to do. As the world is a very very long way away from everyone having a lifestyle like Trump or AlGore, clearly there is a relatively infinite “problem space” in which to employ robots producing goods and services. As robots do more of that portion which they can do, more of what’s left to do falls to people (Amdahl’s Law) and as robots do more, we will demand more and different things in greater quantity (Gustafson’s law).
These two forces working together mean that there is no risk of massive permanent unemployment from automation or robotics. There may well be other things that cause unemployment. (Mostly government rules and brittle labor laws about getting, keeping, and changing professions and licenses) Robots do not guarantee full employment for all. People can be incredibly stupid, especially when it comes politicians and economics. They seem to be a fundamentally bad mix with politicians seeming to think that the ability to sell themselves to ‘the masses’ somehow confers wisdom and understanding of the complex natural process that is a market ecology. Nothing could be further from the truth. Yet robots do not mean unemployment either. If they did, we’d all be unemployed already. The industrial revolution would have done it.
That’s my view on it. Please think about this (and actually read it, even if long) before starting the nay saying and mud tossing.