In my post “The Jobs of Tomorrow: Part 1 of 2,” I wrote that I expect the global economy to transition from the Industrial Age to the “Virtual Age” over the next 25 years, as illustrated in the diagram below. This transition should take place in two parts, the “Remote Control” period (from ~2016-2030) and the “Virtual Public Jobs Programs” period (from ~2031-2040). I will focus on the latter period in this post.
I expect that after experiencing the economic difficulties of the “Remote Control” period, people worldwide will be clamoring for financial assistance from their governments. Technological unemployment is likely to be so widespread by 2030 that many employment relief measures that may seem viable today will be clearly unworkable. For example, mandatory furloughs, reduced hours and minimum wage raises will be pointless if jobs are nearly impossible to get or keep anyway. It will also be too socially costly for governments to abolish technologies that have led to job losses, and mandate that they be replaced with older technologies. Domestic companies that try to revive obsolete technologies will quickly fall behind their international peers, likely forcing them to shed even more jobs.
A large fraction of the population will thus desperately need financial relief, at a time when it’s clear that the available pool of “real world” jobs is both insufficient for everyone and rapidly shrinking. Furthermore, even though the cost of living will probably be lower than today due to technological progress (including the effects of the virtual reality spiral), people will still need sources of income to survive. With private sector sources of income irreversibly disappearing, they will turn to their governments for help.
Governments will then be faced with a very important choice. They will have to choose to either pay their citizens for doing nothing (since jobs will be nearly impossible for them to get and keep), or to pay their citizens for doing something (by providing them with new jobs). The former option, known as a “basic income,” is a relatively old idea that has gained a lot of traction in recent years. Yet, the latter option, hiring citizens to do “public works” projects, proved to be very effective domestically during the Great Depression.
I am confident that public jobs programs will be far more beneficial to future societies than basic income schemes, but only if they are first upgraded to reflect the times, which I discuss below. Once this is done, I expect public jobs programs to be implemented by most countries that are capable of doing so. This will enable them to put many of their unemployed citizens back to work, while simultaneously upgrading their national infrastructures and preparing their populations for the Virtual Age.
Making “Make-Work” Work
During the Great Depression, US President Franklin Delano Roosevelt led the creation of the famous “Works Progress Administration” (WPA), a public jobs program that focused on public works projects.  The WPA built and improved a huge amount of American Industrial Age infrastructure by putting millions of American citizens on the government’s payroll. Although the WPA had to withstand criticisms that some of the jobs it provided were so-called “make-work” jobs , it ended up being hugely successful, and much of the infrastructure it created stands today around the country and is still being used. 
Clearly, the WPA was a phenomenal program for its time. However, today the Industrial Age is rapidly drawing to a close, and we are quickly approaching a period when most physical tasks that people could conceivably be hired to do will be done better and faster by machines. In the 2030s, it will be wasteful and technologically regressive to hire people to do physical labor that machines are easily capable of. It will also be counterproductive to expand global Industrial Age infrastructure at a time when we are likely to have a significant overcapacity of such infrastructure, due to more efficient resource use. Nevertheless, governments should still be able to follow the WPA’s example at this time to relieve unemployment, by providing a completely new class of jobs to their citizens.
Governments should do this by creating public jobs programs that build out and efficiently utilize their countries’ Virtual Age infrastructures. As illustrated in the diagram below, these public jobs programs would be the spiritual descendants of the WPA, upgraded for the 2030s.
To reiterate, Virtual Age socioeconomic activities will be streamlined around reality mimicking activities, rather than reality molding activities, which have characterized the Industrial Age.  To adapt accordingly, societies will thus have to develop advanced virtual reality infrastructures, and automate nearly all supporting “real world” processes. They will also have to repurpose, recycle or retire a considerable amount of legacy Industrial Age infrastructure to support a new focus on networked virtual reality activities.
Although the technical tasks required to build out the foundational components of Virtual Age infrastructure will probably be best left to specialists, efficiently utilizing this infrastructure should enable millions of people to be employed. In fact, former blue, pink, and white collar workers will all be eligible employees if the resulting jobs programs are properly designed.
To accomplish this, governments would fund the creation of massive virtual reality simulations involving huge numbers of people. These simulations would be structured to gain socially beneficial insights on complex social scenarios, which could only be reliably obtained by studying the spontaneous behavior of actual humans in simulated environments. Governments would hire people to:
- Set up the simulations (e.g. scenario design, production design and casting)
- Oversee and maintain the simulations
- Play improvisational roles in the simulations
- Raise public awareness of the simulations
- Scientifically analyze simulation data to gain real world insights
Since many Industrial Age jobs and businesses are likely to be obsolete or largely automated by the 2030s, and there is likely to be a significant overcapacity of Industrial Age infrastructure, Virtual Public Jobs Programs are likely to be one of the only viable public jobs program options. These jobs programs will also have numerous advantages over a basic income, as I illustrate in the table below.
It’s also important to note here that: A basic income is also effectively a type of public jobs program, but one that redundantly pays people to “play themselves” in the “real world,” which they are already doing anyway for free. When viewed in this context, it’s clear that a society that provides a basic income to its citizens gets almost nothing in return for its investment, compared to what it could gain from a well designed public jobs program. Similarly, a citizen that receives a basic income is unlikely to feel a lot of pride about being paid to play a role that s/he is already effortlessly playing for free. WPA Director Harry Hopkins also believed that jobs were more beneficial to the unemployed than handouts of money (known as the “dole”), and phrased his view this way:
Give a man a dole, and you save his body and destroy his spirit. Give him a job and you save both body and spirit. 
Along similar lines, Virtual Public Jobs Programs would provide actual jobs to people instead of a dole, helping them to maintain their “spirit” and dignity. They would also provide societies with highly valuable and otherwise difficult to obtain information that prepares them to make future real world decisions with significantly more confidence than they would otherwise be able to. In addition to this, societies that fund these jobs programs should end up with more advanced virtual reality infrastructures than societies that don’t, due to using their virtual reality infrastructure in more challenging ways. For all of these reasons, I expect these jobs programs to be used to provide employment relief in the 2030s by most countries that have the means to do so.
Fortunately, a lot of the foundation for future Virtual Public Jobs Programs has already been laid by the “Massively Multiplayer Online Role Playing Game” industry, which I expect to play an important role in creating future jobs program simulations. The performing arts and film industries are also likely to be essential to getting these simulations working well.
Yet, a great deal of work lies ahead. In particular, numerous questions about the types of scenarios and behavior that should be permitted in these simulations will need to be thoroughly debated and answered by societies. This will help to avoid some of the unsettling outcomes that occurred in the real life Stanford Prison Experiment .
For a great overview of virtual reality research and the tremendous potential of the medium, I highly recommend the 2012 Jim Blascovich and Jeremy Bailenson book “Infinite Reality.”
Why Machines Won’t Make Humans Obsolete
As genuine machine self-awareness is far from being proven (or even provable) today, humans are likely to be considered much better qualified than machines to play improvisational roles in simulations for some time to come.  Regardless of how advanced machines get, it will also be unknowable whether a machine imitation of a spontaneous human action in a particular situation is what an actual human would have done in that situation. This conundrum is exponentially compounded when considering the collective spontaneous behavior of large numbers of humans from widely varying backgrounds. Since authentic, human oriented results will be the goal of Virtual Public Jobs Program simulations, I expect that human actors will be needed to play reliably authentic human characters for the foreseeable future. 
Nevertheless, it’s evident to me that machines will continue to prove very useful as “nodes” working in conjunction with human “nodes” in a massive global neural net, which is effectively what the internet is today.  This will involve everything from machines performing complex mathematical calculations to machines posing as increasingly convincing extras in public jobs program simulations. The whole created by this combination of human and machine capabilities in a global neural net will prove to be much, much greater than the sum of the individual parts, naturally preventing machines from making humans obsolete. 
What Might a Virtual Public Jobs Program Look Like?
Flight simulators today prepare pilots for dealing with a whole host of potential scenarios, from the most routine conditions to rare crisis situations. Analogously, Virtual Public Jobs Programs should be able to prepare entire societies for dealing with a staggering range of potential real world scenarios. 
As an example of this, let’s imagine that 20 million people are employed to play improvisational acting roles in a publicly sponsored virtual reality simulation, that is scheduled to last a year. Another 20,000 people are also employed to set up, oversee, maintain, publicly promote, and analyze the data from that simulation. All of these employees are required to work during regular working hours only. To facilitate this, the simulation is “paused” at the end of each workday and “unpaused” at the beginning of the next workday.
In this example, the virtual world of the simulation is situated on a fictional continent surrounded by rapidly rising seas. All 20 million actors and actresses play the inhabitants of 10 seaside cities, which are all in a pre-industrial state due to a relative scarcity of fuel sources and minerals. Each actor and actress is assigned a multifaceted role in the simulated society. For example, an actor may simultaneously have daily obligations such as chores or a simulated job, responsibility for other family members, duties associated with a particular social status, and so on. Some roles are designed to be more “key” to the simulation’s outcome than others (like city leaders), but all roles staffed by humans are structured to require a significant amount of improvisational acting.
Before the simulation begins, city leaders are instructed by the simulation’s administrators to save their communities from the rising sea levels by moving them inland. They have a limited set of initially apparent resources available in the simulation to do so, and are given no further guidance on how they should accomplish this, what the conditions are like further inland, or even whether a successful move is possible.
In this scenario, the insights gained from studying spontaneous social responses to rising sea levels in the simulated society could lead to real world insights that save real world money in related scenarios. Because simulation participants will be highly focused on surviving and excelling within the constraints of the simulation, they may be able to come up with solutions to problems that would be ordinarily be overlooked in the real world, which has a different set of circumstances and constraints. To encourage outstanding behavior, simulation employees that perform exceptionally well could also be paid performance bonuses in real life or awarded with other types of recognition.
The potential for this kind of innovation to occur in a simulated scenario was illustrated in the Syfy TV series “Ascension.” In this series, a 350 person crew (later growing to 600 people) is placed on a spaceship in 1963, that simulates departing earth on a century long voyage to a distant planet. Specially designed sets and special effects are used in the story to mislead the crew into believing that the simulated voyage is real, although the spaceship never actually leaves earth. Instead, its crew is monitored over the course of decades by a group of external administrators. Because the spaceship’s crew believes that their situation is real and inescapable, they develop technologies and social structures that are uniquely tailored to their circumstances, and thus quite different from those that develop in parallel in the outside world. 
Along similar lines, a Mars colony simulation (in which participants are of course aware that it’s a simulation) might yield valuable data about how to use limited resources and optimally organize social groups in an environment with very limited resources and resupply options. This would enable a great deal of money to be saved in the event of any future colonization efforts. As Andy Weir’s book “The Martian” fictionally illustrates, even one human in a constrained scenario may be capable of great innovation.  There is no limit to what multiple humans engaged in a simulation that has its own set of constraints might come up with.
Finally, as suggested in Daniel Galouye’s book “Simulacron-3,” simulations could potentially be used to test the effects of product placements on a large simulated population with human characteristics . Rainer Werner Fassbinder’s 1973 movie “World on a Wire” (based on Galouye’s book) also suggests that simulations of this type could be used to reliably and realistically anticipate various economic outcomes.  These possibilities might encourage corporations to sponsor Virtual Public Jobs Program simulations in conjunction with governments, helping to defray their up front costs to societies. Governments may also be able to defray the costs of jobs program investments by producing “simulated reality shows” based on edited portions of ongoing or completed simulations.
References and Additional Notes
 The WPA’s tremendous accomplishments are well documented across a variety of sources. In addition to the Wikipedia link provided, I would recommend the following book as a good introduction: Rose, Nancy E. Put to Work: The WPA and Public Employment in the Great Depression, New York: Monthly Review Press, 2009.
 See for example: Rose, Nancy E. Put to Work: The WPA and Public Employment in the Great Depression, New York: Monthly Review Press, 2009, p.84.
 See for example: http://www.wpatoday.org, which lists WPA accomplishments and provides photographs of many WPA projects that are still in use today.
 Please see my post “Molding Versus Mimicking Reality” for more details on this.
 Referenced from: http://www.pbs.org/wgbh/americanexperience/features/general-article/dustbowl-wpa/, Access date September 7, 2015.
 In addition to the Wikipedia link provided, further information on the Stanford Prison Experiment can be found on the official website. It was also dramatized in the recent movie “The Stanford Prison Experiment” (2015).
 I believe that true machine self-awareness is likely to be rejected by societies for the foreseeable future, as Searle’s “Chinese room” thought experiment basically presents an insurmountable barrier to proof of it. People are assigned responsibility today for machine actions, and this is unlikely to change in the future since machine self-awareness cannot be proved. In this light, “artificial intelligence” claims are likely to be viewed for some time by the majority of people as programmers trying to avoid taking responsibility for the behavior of their machines (i.e. “dog ate my homework” types of excuses).
 Admittedly, the fact that human simulation participants are aware that they are engaged in a simulation might periodically result in over or under exaggerated acting by some participants. However, note that by the 2030s people will have gained considerable experience treating virtual objects and landscapes in shared realities as effectively real. In fact, the Jim Blascovich and Jeremy Bailenson book “Infinite Reality” (HarperCollins ebooks, 2011) gives multiple examples of people responding to virtual situations as if they are effectively real. Therefore it is unlikely to be as big of a stretch as it seems today for people to treat the virtual environment of a public jobs program simulation as also effectively real. Furthermore, I expect the overall quality of improvisational acting to greatly increase as simulations get more and more realistic, and people spend higher and higher percentages of their time immersed in them. The high realism of future simulated environments is likely to play a large role in inspiring authentic performances from simulation participants.
 According to Wikipedia, the closely related term “global brain” was first coined by Peter Russell in his 1982 book “The Global Brain.” Media theorist Marshall McLuhan had also previously popularized the term “global village” in the early 1960s to describe how electronic media was transforming human society. I use the term “global neural net” in this post because I think it better highlights the fact that the internet has enabled both humans and machines to function as “nodes” in a global networked entity, that appears to be evolving towards a focus on advanced problem solving. Note that journalist Paul Mason used similar terminology to describe a modern airliner as “both an intelligent machine and a node on a network,” in his essay “The End of Capitalism Has Begun” (The Guardian, July 17, 2015). This represents a significant shift from the traditional Industrial Age role of both humans and machines as mechanical “parts” of a global economic “machine.” As shown in this post, the global neural net should soon be capable of problem solving skills that are far beyond the realm of simple mathematical calculations, due to its ability to gain insights by realistically simulating alternate human realities.
 It’s very instructive to imagine the global neural net as a networked, semiconscious intelligence that is evolving in its own right, and figuratively striving to gain as much awareness of its surroundings as possible. Since humans function as nodes in this neural net, “real world” human activities that are outside its purview today may be thought of as preventing (or distracting) humans from contributing their full potential to it. However, virtual and augmented reality technologies will dramatically change this situation when they are integrated into the worldwide web, by bringing the bulk of human activities online. So although technology may be thought of as a human tool that is helping humanity to solve Keynes’ “economic problem,” it may also be thought of as a quasi-evolutionary force that is acting to drive every aspect of human activity online, in order to strengthen the capabilities of the global neural net. Viewed in this context, humans are not really competing with machines, even though it may seem so at times. Instead, both humans and machines are being drawn into increasing participation in the global neural net, a giant networked entity that will have far more advanced problem-solving capabilities than any individual human or machine.
 Journalist Paul Mason recently wrote a thought provoking essay about the future impact of technology on society and the economy, titled “The End of Capitalism Has Begun” (The Guardian, July 17, 2015). Mason suggests in his article that societies should work to smooth the transition from the current economy to a future economy with zero marginal costs, a zero carbon energy system and near-zero need for human work. He believes that to facilitate this, it would be ideal to make a highly detailed, open source model of the global economy available to everyone. Using such a model, any individual could then presumably test the economic impact of proposed ideas. In contrast to Mason’s views, I expect first of all that the virtual reality spiral will naturally cause a significant reduction of living costs, and significant increases in the efficiency of resource use. However, no matter how far living costs fall, people will still need some income to cover their expenses, and this income will have to come from somewhere. I also contend that social behavior cannot be authentically modeled unless humans play the role of other humans. It is only when spontaneous human behavior is an integral part of social or economic simulations that societies can have high confidence in their results, and that unexpected forms of innovation are likely to occur. As there are a limited number of humans available to play roles in simulations of this type, and machines are imperfect at doing so, societies will have to prioritize which simulation scenarios are most important to them, thus creating a new source of demand for human employees. Because of this, I expect that public jobs program simulations will continue to add value to societies long after most “real world” work has been automated.
 Besides the Ascension TV show itself, I also used the following sources as references: http://www.syfy.com/ascension/timeline/, Access date September 7, 2015, and http://www.syfy.com/ascension/cast/1, Access date September 7, 2015.
 Weir, Andy. The Martian, New York: Crown Publishing Group, 2014 (also “Originally self-published as an ebook in 2011,” according to the front matter)
 Galouye, Daniel. Simulacron-3, Rockville, MD: Phoenix Pick, 1964.
 The 1973 Rainer Werner Fassbinder film “World on a Wire,” based on Daniel Galouye’s book Simulacron-3, mentions the possibility of using simulations with realistic characters to anticipate economic outcomes during a press conference in the movie.