Still optimistic about the future of jobs in a bot driven world

Thanks to everyone who contributed to my optimistic discussion about the future of jobs in a bot driven world published earlier this month.

The quality of thinking and the passion of this readership is encouraging. In this article, I'd like to address some of the many thoughtful comments that were left. Before doing so, it's important to draw attention to some of the underpinnings of my thinking which you can find here and here. They provide both a historical and economic context as to why I am optimistic, despite the many dire predictions that have, in my view, raised unn-necessary alarm.

As in all these kinds of discussion, hindsight is something we don't yet have but for now, it is useful to talk to the issues raised.

Automation creates leisure time

With the advent of the steam engine, all kinds of leisure time was predicted. this was because most people were in agriculture and they figured they would be out of work now that their jobs were automated. I imagined there were even some who went looking for handouts. But as usually happens in history, the advent of a new tool, if available to the people, produces all kinds of new industries. the steam engine (and similar technologies) lead to the rise in manufacturing and the exodus to large cities.

Increasing leisure time has been happening ever since people began living together in communities.

As specialization happened most people could work less. The big leap was from hunter-gatherer to farming. Recent research suggests that modern hunter-gatherers in remote areas like the Amazon can live by working about 20 hours per week. The rest is leisure.

When we formed larger societies we discovered we needed to work a bit more. Farming was a practically non-stop effort at least for part of the year. But there was a trade off of either living in the wilderness with a lifespan of about 30 years if you were a man and maybe less if you were a woman (people used to die in child birth at alarming rates) OR working a bit more to provide a more secure lifestyle. Our ancestors chose the latter.

Where did the leisure go? Better food and nutrition take time, so do school, religion, and community commitments like being in a militia that defended the homestead. But there was also art especially performance which we condense into TV and movies today. We spend hours in front of screens daily often for fun. Leisure has always been there and it has always gobbled up free time leaving us to wonder where it went.

Whether we consider that a net 'good' in modern America is moot at a time when the Protestant work ethic is being questioned but the fact remains we should expect technology advances to offer us the opportunity and choice to enjoy a greater degree of leisure.

Technology is hoarded

The important lesson to learn from this is not that technology will just magically produce new jobs like it always has. The trend for a new tool is for it to be hoarded. The important lesson to learn here is that we should champion those companies that make these tools available at reasonable prices to the rest of us.

In a long economic wave invention is first expensive and bought by early adopters who are the only ones that can afford it. The balance of the cycle is all about commoditization and bringing innovation to the masses. That’s where we are today; we’re in the efficiency-commoditization part of the cycle.

The good news is that information and its technologies are nearly free today. The bad news is that it’s increasingly difficult to generate revenue from many technologies. A 1950s era mainframe cost millions of dollars (just a MB of memory cost a million dollars in 1955). Today we carry around much more processing power including in the phone service we subscribe to. That’s the commoditization cycle.

The combination of Moore's Law, the rocket ship pace of smartphone adoption coupled with a ubiquity of readily available content and comment has heightened our collective awareness of what is on offer and, in one sense, our increased need for instant gratification only serves to fuel our desire to see more progress in the shortest time span.

Drone is not automation

Your analysis is incomplete and your drone example is not even automation. Automation in that case would be having a drone that knows where it should go all by itself without the need of any pilot. CGP grey explains very nicely Why this automation is different?

A drone today is first generation automation and it’s important to understand the historical context for how automation evolves.

As an example, in the early years of the Industrial Revolution a Spinning Jenny, which was a machine (automation) that made cotton thread, produced thread for 12 spindles and it was operated by one person, an adult male. Fifty years later it was common to find Spinning Jennies that served over 1,000 spindles at once and these were managed by 12 year-olds.

The job was no longer enough to pay a head of household and these machines made thread especially cheap. This is what happens in any economic cycle.

A drone is not the best example but it is automation in the sense that it economizes warfare in ways that make it possible to prosecute multiple battles vastly cheaper than conventional warfare and at no cost of life to those who have the technology. Not pleasant to think about but then there are fresh uses for drones in cinematography that open up opportunities to experience the world through augmented and virtual reality that we are only just discovering.

A better example might be the story below about antibiotics as a form of commoditization and automation.

Automation is taking over unskilled jobs

The main trouble is that automation is expanding to take over all unskilled jobs. The time is coming when 4 years of university is the bare minimum education a human can have to be employable. This means that all those people that barely got through high school are out of luck, and even those well educated now will have to compete for what work remains.

That mostly appears to be true. But the article makes the point that most of those jobs fall under the 3D’s of dull, dirty, and dangerous—thanks to Vinnie Mirchandani, author of “Silicon Collar— An Optimistic Perspective On Humans, Machines And Jobs” Automation is making new jobs though there is always a lag time between the turning of a cycle and when it can generate enough jobs to drive the economy. See my story about Arthur C. Clark below.

As far as a four-year degree is concerned there’s room for doubt. There’s been a great deal of education inflation over the last generation.

Jobs that never needed college now do as we have professionalized things that simply don’t need that status. I went to cooking school in the 1990s on weekends for fun. It was a two-year program that I didn’t finish and it focused on what you needed to know to run a commercial kitchen. Today that’s a four-year degree padded with things that might be useful but that frankly could be picked up either in a first job or more properly in high school.

This over professionalization places a barrier between jobs and capable people. We need to re-examine what college is designed to impart.

Serving many with one in the professional world

The problems continue: The higher education occupations typically serve many people at once. A doctor has many patients, an engineer works on many products that sell many copies each, etc.

It has been only since WW2 that a trip to the doctor was more likely to help than harm, thanks to advances in antibiotics, X-rays and medical science. Doctors used to be more one on one with patients largely because they didn’t know what they were looking at in a disease or set of symptoms.

Doctors spent centuries collecting data until they were able to crunch it and come up with known disease states and reliable therapies. I remember speaking with one MD who was in his 90s at the time and he told me when antibiotics first appeared they were given by injection 3-4 times a day. The powdered penicillin had to be mixed in a vial with saline solution and then administered. So if a patient needed a 10-day course of treatment that meant 30-40 house calls and office visits! It was a major commitment.

You don’t think of a tablet or capsule of antibiotic as automation, but it is. It saves time and human (doctor) labor so that the doctor can do more value added things. And, yes, a doctor’s dance card has back filled very nicely since that particular automation. We need automation to do exactly what it does so that we can advance society.

We won’t be able to do those new jobs because the math is too hard

Only one in ten people are really genetically and environmentally optimized to be STEM (Science, Technology, Engineering, and Mathematics) students and graduates. So, claiming technology is going to generate new STEM jobs is no consolation for the 9 of ten people who fall outside the STEM optimum.

The point of automation is to hide complexity.

Recall Arthur C. Clark’s encomium that:

Any sufficiently advanced technology is indistinguishable from magic.

Owning and driving a car was once a major undertaking requiring the owner to be able to effect many repairs and to have a working knowledge of how the machine operated. Today, not so much unless you happen to be someone who enjoys tinkering with cars. You get in and drive.

Technology is becoming magic as it commoditizes. We code software way less than we once did to a point where hardcore machine coding is fading into the background, allowing people to express their creative and design skills while the machines generate working code. I no longer know the ins and outs of my operating system because I don’t have to and that’s especially true for my iPad. It’s the way of the world.

By the same measure, we assemble websites like diginomica from preprogrammed components rather than sweating hours over scripts, PHP and database table configurations. And make no mistake, enterprise vendors are thinking the same way, looking to make it much easier for business to assemble attractive online presences than would have been the case just five years ago.

As for STEM jobs, not everyone is cut out to do them, granted. But not everyone is cut out to write hit songs or play pro ball. You make picks and chase your dreams. The expansion of technology creates new niches that we can disappear into for a career. The glass is half full. Really, and I'd argue that we not only have plenty of choice but we can choose to do those things that fulfill us as individuals rather than being tied to rote, mindless tasks. Which do you prefer?

I am naïve

I think Denis Pombraint’s optimistic view comes from the observation that since the first Industrial Revolution, displaced jobs have been replaced by new jobs. So it seems to make sense that there will be future jobs for people able to achieve what intelligent automation (IA) cannot. I think this naive.

Yes, I probably am naïve, but not about this.

Going back to the idea that this has happened before but that most people don’t know or recall, this is a cyclical pattern called a K-wave. Also above, automation is opening up new niches.

The hallmarks of K-waves are that they are initially inflationary, employ a lot of people, make huge profits, and generally drive the economy for about 50 years. However, the second half of the wave is all about consolidating gains from the first part.

It is studded with economizing, efficiency, and process improvements to wring the last bit of profits from the initial invention. Guess where we are in all that? Yup, the second half.

This wave is technically called The Age of Information and Telecommunication and it is commoditizing rapidly. IT and Telecoms won’t go away just as steel and petroleum and cars haven’t but they have been commoditized to the point that they don’t drive the economy in perceptible ways. Steel is made where it can be made cheapest, textiles, (and computers) too. In each case and over almost two centuries we’ve watched these industries travel around the world.

Take cotton for example. In the 19th century a high proportion of earth’s cotton supply came from North America where it could be grown almost for free thanks to slave labor. The cotton exchange in Liverpool, England managed the global supply chain connected by undersea telegraph cable. It was the first instance of globalization. When American cotton became too expensive capital took flight and moved farming it to less developed parts of the world.

The same goes for textile manufacturing. Initially most cotton went to the UK's north or to New England factories to be turned into cloth. When labor became expensive, the industry migrated around the world from New England to the American South, from Lancashire England to India and Bangladesh to the Far East and places like China and Vietnam today. The whole paradigm of cotton and textiles commoditized and does not drive the western economy any more though we still need its output.

That does leave open the question of how far automation can go but I suspect that in this type of industry, we're witnessing an endgame of sorts where the lowest cost labor economy gives way to near full scale automation.

Not enough jobs are being generated

Hey nice positive article, but I think you are missing the math. A YouTube person who has an economics degree pointed out the problem. It’s not that all the jobs will suddenly disappear, it’s that very large employers of low skill jobs will disappear quickly and cause economic problems for a lot of reasons. It is also a problem that the amount of new jobs are much less than the ones automated so it’s not a zero sum gain.

True, not enough new jobs are appearing mostly because we’re looking at old paradigm jobs, which are disappearing. Even so, the census data alone suggests the US needs to create 150,000 jobs a month for the foreseeable future simply to cope with population expansion. Some doubt this number but the fact remains that the jobs market is continuing to expand. Many of those jobs will be low paying and will strain current HR systems as we move to a more contract labor based approach to workforce management. But this change has already been anticipated and you can expect to see this play out over the next 5-10 years.

Alongside these numbers, Vinnie Mirchandani’s book suggests that as many as 9 million jobs can be found across some 800 plus categories.

There are plenty of low skill jobs around, but they don’t pay very well, that’s the problem. People won’t do a job that doesn’t pay a living wage and why should they? If the scuttlebutt from Trump Tower is accurate we’re getting ready to spend a trillion dollars in the U.S. on infrastructure. Not only will this generate short-term construction jobs, but also the investment in productive capacity will have a positive feedback as this type of investment consistently creates new permanent jobs of the kind people are hoping for.

In the late 20th and early 21st centuries, the Commonwealth of Massachusetts engaged in a mega construction project that cost about $14 billion and lasted roughly 20 years. It was to build a tunnel system to complete U.S. Route 90 from Boston to Logan airport, a distance of a few miles.

The project created new construction jobs that are now gone but it improved the productive capacity of the region where millions of people access the roads, bridges, and tunnels and do productive work in manufacturing, science, medicine, and technology.

A couple of years ago General Electric moved its headquarters from Connecticut to Boston bringing with it many new jobs. GE’s focus on tech and building the operating system of the manufacturing future is generating lots of those jobs (not all of them in Boston) and they are net-new jobs. How many? I don’t know. That’s something that flies under the radar but it’s not just tech jobs we’re talking about it’s hotel and restaurant jobs, public safety jobs, and many others.

It's useful to note that when technology comes to town, it is estimated that four more jobs are created in support of each high technology worker.

My final thoughts (for now)

The one thing we didn’t touch on in any depth is the impact of population growth. When the Information and Telecommunications Age started in about 1971 there were roughly 3 billion people on the planet. The population has since more than doubled creating a major challenge not only for employment but also for resources.

Back then the planet used about 25 million barrels of petroleum per day, today the U.S. uses about 20 million bbls and the planet just under 80 million bbls. Finding jobs, food, and other resources for this growing population is a big ask and it’s getting harder.

With so many people looking for work and an economic framework that favors rent seeking, capital has the upper hand in hiring the lowest cost work force it can. The obvious conclusion is that we need to stabilize population, a thought that no one wants to contemplate. But scarce resources on this level is simply a negative feedback loop on unconstrained population growth. That’s a weighty thought and I draw no conclusions; it’s a matter of personal ethics, but it’s out there.

In 1970 the Club of Rome published a report titled, “Limits to Growth.” You can pick up on updated version on Amazon that adds 30 years of data. “Limits” predicted some of what we’re seeing today. It was largely ignored and ridiculed but it remains valid and top of mind for some people like me.