uppose Lake Michigan is somehow emptied. How long would it take to fill the lake with water if you can only add one ounce of water and double it every 18 months? The answer is 85 years. Actually, almost all of the water volume starts expanding only in the last five years. After 60 years, the bottom of the lake only gets a little damp; after 70 years, only a few puddles exist; after 80 years, there are only 40 feet of water, and then suddenly the lake is filled in the last five years.
This was the example Kevin Drum used in 2013 to explain exponential growth. As human beings, we are not used to understanding exponential changes because we experience time in a linear way. We tend to rely on the changes we have seen in the past to predict the future, assuming a constant rate of change. But this example helps us capture this idea—an extremely long period during which the observable progress is negligible, followed by an explosion.
It sounds like a fantasy, but it is exactly how exponential technology works.
Exponential technology has two attributes. First, it can double its performance or capacity (or cut its cost in half) in each period, based on the famous Moore’s Law—computing power doubles around every 18 months. Artificial intelligence (AI), augmented and virtual reality (AR, VR), biotech, robotics, and autonomous vehicles are all examples of exponential technologies.
Second, it has to be in a development phase where a cost-benefit calculation allows it to solve today’s problems in previously unimaginable ways. Think about drones. They used to be so expensive and only the military could afford them. Now everyone can have them for a few hundreds of dollars.
Given their huge potential, exponential technologies bring great hope in helping us confront the world’s most pressing challenges. Experts believe that the possibility of developing sustainable solutions becomes more likely once we use two or more of these exponential technologies together to attack a persistent challenge.
What if exponential technology grows at a pace we cannot predict or master? Will its application cause more unsolvable problems? Probably yes, and so did every other innovation that facilitated human development. If the application of exponential technologies will come sooner or later, we should at least try to understand its potential consequences that we can think of today.
Change in Labor
People are worried that robots will take their jobs. According to McKinsey Global Institute research, even though less than 5% of jobs can be fully replaced, about 60% of occupations have at least 30% of activities that have automation potential.
Undoubtedly, the way people work in the future will be different. Workers will work more closely with technology and they will be hired to do the activities that are complementary to machines. The jobs that require repeatable physical actions, elementary calculation, and data collection—basically any activity that can be programmed into an “if-then-do” function—will no longer exist.
It sounds like a huge unemployment alarm. However, if you agree that we are experiencing the fourth industrial revolution, you might be less concerned. After all, industrial revolutions throughout history are characterized by changes in labor—the way we work—and this one should be of no exception. As Mark Michaelis reminded us in 2019, after the technology that triggered each revolution was democratized, labor improved, as all the workers moved to better jobs in the end.
Will this revolution be any different?
Change in Industries
To answer this question, let’s first turn to some of the huge disruptions that might take place in a few industries.
Retail. How often have you used grocery delivery services to keep social distancing in the past few months? If drone delivery is faster and cheaper, how often will you go to retail stores again? As Creative HQ predicted, retail might go the way of banks, using fewer storefronts, which means there will be fewer employment opportunities for security, cashier, salesmen, janitors, and so on.
Business. Andrew Fursman, an expert in quantum computing and network and computing systems, predicts that we will be able to monitor and analyze every financial transaction in the world simultaneously by 2027. We might not need accountants anymore. Also, think about financial analysts. How can they compete with machines with strong predictive analytic capabilities that can access all the real-time information?
Medical. What about the medical professions that are harder to replace? Even the McKinsey report ranked healthcare with very low automation potential. Kevin Drum, the designer of our Lake Michigan refilling project, is less optimistic. In his article, he mentioned a therapy simulation program named Eliza, created by an MIT computer scientist Joseph Weizenbaum nearly 50 years ago. It was only made with a keyboard and an old teletype terminal, but people liked talking with Eliza, because Eliza was endlessly patient and extremely interested in their problems. With rapid technology development in the past 50 years and people’s growing dependence on technology, artificial companionship is becoming a new normal, said Sherry Turkle, an MIT professor who studies human-technology interaction.
Change in National Income Distribution: Capital vs. Labor
It seems that exponential technologies might reduce more employment than we expected. One old concern in economics is then raised— “capital-biased technological change,” meaning that a larger income distribution will shift from workers to capital owners. While labor’s income includes wages, salaries, bonuses, and things like insurance benefits, capital income refers to interests, dividends, investment gains, and all other capital gains.
The data already shows this notable income distribution shift. Admittedly, as Robert Lawrence from Peterson Institute for International Economics (PIIE) points out, many other possible reasons are behind this change such as globalization, aging population, and a weak bargaining power of labor. However, automation is an extremely important one because of the substitutional effect between capital and labor. When exponential technologies come into play, this substitutional effect will only be stronger. Of course, some early-stage technology can complement human jobs, but eventually, machines still perform these combined missions well, given their exponentially augmented capacities.
American economist, Paul Krugman expressed his worries about this income inequality caused by the “Rise of Robots.” He believed that all the conventional ways to reduce inequality will not work in this capital-dominant future. Neither better education nor an “opportunity society” will do much if the most valuable assets are the ones inherited from your parents.
Another economist Noah Smith also worried about what this huge income inequality means for labors, and he gave some interesting suggestions to redistribute wealth and income against machines in the future. For now, everyone is born with an endowment of labor and we can trade it for income. However, when what we are endowed with is worth less and less, and when capital is taking larger shares of income, why not also give people some endowment of capital? For example, when each individual turns 18, the government can buy him/her a diversified portfolio of equity or give him/her some robots, which can act as an insurance policy for each human worker.
So, back to our earlier question: will this revolution be any different? Probably yes. Of course, we will experience a change of labor, as we did in every past revolution. However, due to the exponential growth of these technologies, a lot of structures—such as economic, business, and educational ones—are likely to be disrupted and changed, developing into directions that are hard to foresee or predict with linear thinking.
This unpredictability also poses great challenges for governance and raises a lot of ethical concerns.
First, exponential technologies will make our already lagging-behind regulation-making process more obsolete. Regulations can hardly be useful if innovation process always outruns the bureaucracy. It takes more time and effort to design and agree on protocols when all countries’ cooperation is needed to address this global concern. Take cybersecurity agreements as an example. It takes rounds and rounds of multilateral meetings to settle on what each norm means in different languages, and this will become more complicated when exponential technologies are in their explosion phase.
Second, emerging technologies are likely creating some ethical questions, which have not gotten enough attention yet. Experts are worried that too few organizations have ethicists on their team, which might easily blur the boundary between the beneficial and the destructive. Genetic engineering is an example, and there are more to come. When technology gives humans increasing—exponential—power, where/what is the red line that humans should not cross?
Despite the difficulties in governance and ethical dilemmas created by exponential technologies, humans should be actively trying to understand and adapt to this technology-driven future.
On an individual level, humans should focus on developing creativity-oriented future-proof skill sets. In a 2013 white paper titled “Dancing with Robots: Human Skills for Computerized Work,” Richard Murnane and Frank Levy summarized three things humans are still better than robots: solving unstructured problems, working with new information, and doing non-routine manual work. A World Economic Forum report in 2016 also identified a set of “21st century skills,” which includes foundational literacies (reading, writing, science), competencies (critical thinking, creativity, communication, and collaboration), and character qualities (curiosity, persistency, adaptability, and leadership).
On an organizational or national level, exponential mindsets should be developed to embrace the possibilities and future changes. Organizational structures at any level should be kept flexible enough to allow innovation teams at the edges to make breakthroughs. Also, since technology operates beyond boundaries, both public-private partnerships (PPP) within a nation and multilateral cooperation across countries should be considered to keep practices in order, exchange ideas, and foster innovation.
We are exploring a new era and our past experiences are not enough to guide us anymore. Exponential technologies can easily move from a deceptively slow pace of development to a disruptively fast pace and we cannot afford to take that lightly.
a global affairs media network
Moving from Bold Thinking to Action
July 27, 2020
S
uppose Lake Michigan is somehow emptied. How long would it take to fill the lake with water if you can only add one ounce of water and double it every 18 months? The answer is 85 years. Actually, almost all of the water volume starts expanding only in the last five years. After 60 years, the bottom of the lake only gets a little damp; after 70 years, only a few puddles exist; after 80 years, there are only 40 feet of water, and then suddenly the lake is filled in the last five years.
This was the example Kevin Drum used in 2013 to explain exponential growth. As human beings, we are not used to understanding exponential changes because we experience time in a linear way. We tend to rely on the changes we have seen in the past to predict the future, assuming a constant rate of change. But this example helps us capture this idea—an extremely long period during which the observable progress is negligible, followed by an explosion.
It sounds like a fantasy, but it is exactly how exponential technology works.
Exponential technology has two attributes. First, it can double its performance or capacity (or cut its cost in half) in each period, based on the famous Moore’s Law—computing power doubles around every 18 months. Artificial intelligence (AI), augmented and virtual reality (AR, VR), biotech, robotics, and autonomous vehicles are all examples of exponential technologies.
Second, it has to be in a development phase where a cost-benefit calculation allows it to solve today’s problems in previously unimaginable ways. Think about drones. They used to be so expensive and only the military could afford them. Now everyone can have them for a few hundreds of dollars.
Given their huge potential, exponential technologies bring great hope in helping us confront the world’s most pressing challenges. Experts believe that the possibility of developing sustainable solutions becomes more likely once we use two or more of these exponential technologies together to attack a persistent challenge.
What if exponential technology grows at a pace we cannot predict or master? Will its application cause more unsolvable problems? Probably yes, and so did every other innovation that facilitated human development. If the application of exponential technologies will come sooner or later, we should at least try to understand its potential consequences that we can think of today.
Change in Labor
People are worried that robots will take their jobs. According to McKinsey Global Institute research, even though less than 5% of jobs can be fully replaced, about 60% of occupations have at least 30% of activities that have automation potential.
Undoubtedly, the way people work in the future will be different. Workers will work more closely with technology and they will be hired to do the activities that are complementary to machines. The jobs that require repeatable physical actions, elementary calculation, and data collection—basically any activity that can be programmed into an “if-then-do” function—will no longer exist.
It sounds like a huge unemployment alarm. However, if you agree that we are experiencing the fourth industrial revolution, you might be less concerned. After all, industrial revolutions throughout history are characterized by changes in labor—the way we work—and this one should be of no exception. As Mark Michaelis reminded us in 2019, after the technology that triggered each revolution was democratized, labor improved, as all the workers moved to better jobs in the end.
Will this revolution be any different?
Change in Industries
To answer this question, let’s first turn to some of the huge disruptions that might take place in a few industries.
Retail. How often have you used grocery delivery services to keep social distancing in the past few months? If drone delivery is faster and cheaper, how often will you go to retail stores again? As Creative HQ predicted, retail might go the way of banks, using fewer storefronts, which means there will be fewer employment opportunities for security, cashier, salesmen, janitors, and so on.
Business. Andrew Fursman, an expert in quantum computing and network and computing systems, predicts that we will be able to monitor and analyze every financial transaction in the world simultaneously by 2027. We might not need accountants anymore. Also, think about financial analysts. How can they compete with machines with strong predictive analytic capabilities that can access all the real-time information?
Medical. What about the medical professions that are harder to replace? Even the McKinsey report ranked healthcare with very low automation potential. Kevin Drum, the designer of our Lake Michigan refilling project, is less optimistic. In his article, he mentioned a therapy simulation program named Eliza, created by an MIT computer scientist Joseph Weizenbaum nearly 50 years ago. It was only made with a keyboard and an old teletype terminal, but people liked talking with Eliza, because Eliza was endlessly patient and extremely interested in their problems. With rapid technology development in the past 50 years and people’s growing dependence on technology, artificial companionship is becoming a new normal, said Sherry Turkle, an MIT professor who studies human-technology interaction.
Change in National Income Distribution: Capital vs. Labor
It seems that exponential technologies might reduce more employment than we expected. One old concern in economics is then raised— “capital-biased technological change,” meaning that a larger income distribution will shift from workers to capital owners. While labor’s income includes wages, salaries, bonuses, and things like insurance benefits, capital income refers to interests, dividends, investment gains, and all other capital gains.
The data already shows this notable income distribution shift. Admittedly, as Robert Lawrence from Peterson Institute for International Economics (PIIE) points out, many other possible reasons are behind this change such as globalization, aging population, and a weak bargaining power of labor. However, automation is an extremely important one because of the substitutional effect between capital and labor. When exponential technologies come into play, this substitutional effect will only be stronger. Of course, some early-stage technology can complement human jobs, but eventually, machines still perform these combined missions well, given their exponentially augmented capacities.
American economist, Paul Krugman expressed his worries about this income inequality caused by the “Rise of Robots.” He believed that all the conventional ways to reduce inequality will not work in this capital-dominant future. Neither better education nor an “opportunity society” will do much if the most valuable assets are the ones inherited from your parents.
Another economist Noah Smith also worried about what this huge income inequality means for labors, and he gave some interesting suggestions to redistribute wealth and income against machines in the future. For now, everyone is born with an endowment of labor and we can trade it for income. However, when what we are endowed with is worth less and less, and when capital is taking larger shares of income, why not also give people some endowment of capital? For example, when each individual turns 18, the government can buy him/her a diversified portfolio of equity or give him/her some robots, which can act as an insurance policy for each human worker.
So, back to our earlier question: will this revolution be any different? Probably yes. Of course, we will experience a change of labor, as we did in every past revolution. However, due to the exponential growth of these technologies, a lot of structures—such as economic, business, and educational ones—are likely to be disrupted and changed, developing into directions that are hard to foresee or predict with linear thinking.
This unpredictability also poses great challenges for governance and raises a lot of ethical concerns.
First, exponential technologies will make our already lagging-behind regulation-making process more obsolete. Regulations can hardly be useful if innovation process always outruns the bureaucracy. It takes more time and effort to design and agree on protocols when all countries’ cooperation is needed to address this global concern. Take cybersecurity agreements as an example. It takes rounds and rounds of multilateral meetings to settle on what each norm means in different languages, and this will become more complicated when exponential technologies are in their explosion phase.
Second, emerging technologies are likely creating some ethical questions, which have not gotten enough attention yet. Experts are worried that too few organizations have ethicists on their team, which might easily blur the boundary between the beneficial and the destructive. Genetic engineering is an example, and there are more to come. When technology gives humans increasing—exponential—power, where/what is the red line that humans should not cross?
Despite the difficulties in governance and ethical dilemmas created by exponential technologies, humans should be actively trying to understand and adapt to this technology-driven future.
On an individual level, humans should focus on developing creativity-oriented future-proof skill sets. In a 2013 white paper titled “Dancing with Robots: Human Skills for Computerized Work,” Richard Murnane and Frank Levy summarized three things humans are still better than robots: solving unstructured problems, working with new information, and doing non-routine manual work. A World Economic Forum report in 2016 also identified a set of “21st century skills,” which includes foundational literacies (reading, writing, science), competencies (critical thinking, creativity, communication, and collaboration), and character qualities (curiosity, persistency, adaptability, and leadership).
On an organizational or national level, exponential mindsets should be developed to embrace the possibilities and future changes. Organizational structures at any level should be kept flexible enough to allow innovation teams at the edges to make breakthroughs. Also, since technology operates beyond boundaries, both public-private partnerships (PPP) within a nation and multilateral cooperation across countries should be considered to keep practices in order, exchange ideas, and foster innovation.
We are exploring a new era and our past experiences are not enough to guide us anymore. Exponential technologies can easily move from a deceptively slow pace of development to a disruptively fast pace and we cannot afford to take that lightly.
