reflection #3

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 1

Fourth Industrial Revolution

The phrase “Fourth Industrial Revolution” is increasingly being used in recent years. To understand the phrase, let’s review what we’ve learned.

• Early industrialization was from the late 1700s to the mid-1800s – the First Industrial Revolution.

• Later industrialization was from the late 1800s to the mid-1900s – the Second Industrial Revolution.

• The Post-Industrial stage started in the late 1900s – the Third Industrial Revolution.

o Remember, “Post-Industrial” does not mean manufacturing fewer industrial goods. It means manufacturing industrial goods with fewer industrial workers.

o Employment shifts from jobs in manufacturing to jobs in information-based services and technology. This shift often contributes to inequalities of income.

• The phrase Fourth Industrial Revolution refers to developments in the 21st century, especially starting in the 2010s.

o These developments include Artificial Intelligence (AI), 5G broadband telecommunications, and Quantum Computing.

Let’s look at these developments. As we do so, we’ll return to important themes we’ve discussed in earlier phases of industrialization such as creative destruction and energy transitions.

Artificial Intelligence (AI) The term “Artificial Intelligence” refers to computers processing information in ways that resemble human intelligence. More specifically, Artificial Intelligence means computer programs, computer systems, and computer-controlled robotics processing information in ways that resemble the intellectual processes of humans.

• That’s a complex statement.

• Read it again – slowly.

• Stop and think through what the statement means. Now let’s begin to think about AI by using a simple example – chess. We probably all know that we can download a computer program which allows us to play chess against the program. Chess requires an intellectual process. Each player has to process and respond to the moves of the other player. A computer program playing chess is a simple example of AI. The program’s processing of information involves a basic level of intelligence – the ability to memorize individual items (chess pieces) and procedures (the rules of chess). This ability to memorize items and procedures is called “Rote Learning.”

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 2

• Chess, of course, is not a simple game. It involves computing hundreds, thousands, and even hundreds of thousands of possible moves for numerous pieces which each move according to different rules.

• But computers have long been able to process millions of pieces of information to complete complex tasks – like solving difficult mathematical problems.

• Processing all the possible moves in chess is a complex task. But the kind of computer learning involved is the simple level of intelligence called Rote Learning – the ability to memorize and apply rules.

Let’s now consider a higher kind of intelligence than Rote Learning. Let’s consider the intelligence called “Generalization.” Remember, when we talk about “intelligence” we are talking about an intellectual process. Generalization involves the process of drawing inferences – using evidence and reasoning to draw conclusions, and then testing those conclusions through trial and error.

• Here’s a simple example of Generalization: I drink two cups of coffee in a day. That night, I wake up at 3 a.m. and can’t get back to sleep.

• I use my intelligence to generalize – I use evidence and reasoning to draw a conclusion. My evidence is that I had two cups of coffee and I woke up at 3 a.m. My intelligence generalizes by connecting these two pieces of evidence and drawing the following conclusion: The amount of caffeine in two cups of coffee prevents me from sleeping well.

• But maybe my conclusion is wrong. Maybe I connected two pieces of evidence which are not really connected. Maybe I woke up at 3 a.m. for another reason, such as feeling anxious about a biology test the next day. Maybe it was anxiety, not caffeine, which interrupted my sleep. I take this a step further with trial and error, meaning I repeat the drinking two cups of coffee for several days to see if I keep waking up at 3 a.m.

Notice how Generalization is a higher kind of learning than Rote Learning. Generalization is not just memorizing rules and procedures (Rote Learning). It is using the power of reasoning to examine evidence, to draw conclusions by connecting different pieces of evidence, and then to test those conclusions through trial and error. But can a computer do this higher kind of learning called Generalization? Can a computer do the intellectual process of reasoning? This would mean it could use reason to

1. examine evidence

2. draw conclusions by connecting different pieces of evidence which humans did not program the computer to connect

3. test those conclusions through trial and error. The key here is connecting evidence “which humans did not program the computer to connect.” Let’s highlight this point by using the example of medical diagnoses.

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 3

A computer can be programmed to connect three pieces of medical information to suggest a diagnosis. Someone writes a program which “reasons” that if a patient has a particular genetic marker along with two specific symptoms, then there’s an 85% chance that patient has or will develop a particular disease. Such a program connects evidence (three pieces of medial information) in order to draw a conclusion (85% chance of having a disease). But this level of intelligence is Rote Learning. The computer is programmed to make the connections between the evidence and the conclusion. Computers can perform this level of AI. But can computers perform the higher level of intelligence called Generalization? This would mean the computer connects pieces of medical information (evidence) which it was not programmed to connect, and then suggests a diagnosis (a conclusion). To think this through, let’s consider more than just one patient. Let’s consider millions of patients. If AI could perform the higher kind of learning called Generalization – the intellectual process of reasoning – it might mean something like this:

• The medical information and diagnoses of millions of patients could be put into a program.

• The program could analyze the information and see connections between genetic markers, symptoms, and diseases which it was not specifically programmed to find.

• The program then repeatedly tests its conclusions by drawing on continuously updated genetic histories, symptoms, and diseases of millions of people.

• Further information could be added beyond genetics, symptoms, and diseases, information such as environmental conditions, dietary habits, etc. Such information would require the program continuously to test and revise its conclusions because environmental conditions might affect certain genetic markers differently.

This scenario highlights the current limits of Artificial Intelligence, meaning it’s unclear how well AI can perform this level of Generalization. Remember, the goal of AI is to process information in ways that resemble human reasoning – examining and connecting evidence leading to a conclusion which is further tested through trial and error. Notice what this level of human reasoning involves. It involves flexibility and adaptation. Trial and error mean always testing and challenging the conclusions one draws. Trial and error thus require the intellectual flexibility to adapt one’s thinking as more evidence arises – always being open (flexible) to more evidence and thus the need to revise (adapt) one’s conclusions. This kind of flexibility and adaptation add extra layers of complexity and possibilities to the reasoning process. In other words, human reasoning (the ability to Generalize) involves these four characteristics – flexibility, adaptation, complexity, and unknown possibilities. Let’s use some examples of Artificial Intelligence to consider whether AI can perform these four characteristics of human reasoning. We’ll focus on the use of AI in manufacturing, mining, and language recognition.

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 4

Let’s start with manufacturing. Consider the process of manufacturing a car. The process includes thousands of steps. One step is to paint the body of the car. Here are two pictures of robots painting a car.

Robots have been painting cars this way for decades. The robots (computers) are programmed to open doors and hoods, and to spray paint in specific ways. But notice that computers painting cars are not doing the higher level of reasoning called Generalization. They are not examining and connecting evidence leading to a conclusion. Rather, the computers are programmed to perform specific tasks like a computer playing chess. Let’s go further by considering the use of AI in mining – i.e., mining minerals from the ground. Mining has been crucial since the beginning of industrialization such as the mining of iron and coal to make and operate the steam engine. Mining remains crucial in the 21st century such as the mining of lithium for lithium-ion batteries. In the pictures below, we see the use of AI in 21st century mining. Technicians use virtual reality to control robots which do the mining:

These pictures suggest that this level of AI is similar to computers painting cars. It does not require computers to do the higher-level reasoning called Generalization. Rather, it simply requires computer networks (such as 5G networks) to allow technicians to control and manipulate robots from a distance. It’s impressive, but it’s still the technicians (the humans) doing the reasoning, not the robots themselves.

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 5

To push AI further involves computers doing reasoning. It’s unclear how much or how complex reasoning computers can do, but current attempts to push the limits of AI involve computers doing reasoning like problem solving. Consider the following scenario: A computer is programmed to achieve a goal – to mine a particular mining shaft. The computer is also programmed to perform certain actions – drill, move left, move right, move forward or back, etc. Problem solving involves the computer selecting the best order of actions to reach the goal – when to move left, when to move right, when to drill, for how long, etc. If the actions are performed in the best order, the process is more efficient – more is done in less time (productivity increases). If the actions are performed in the wrong order, the process is less efficient (productivity decreases). Let’s think this through:

• If the technicians controlling the robots make all these decisions – when to move the drill right or left, where to drill, and for how long – then the level of AI is relatively simple, like robots painting cars.

• But if the computer itself can choose actions from a range of possible actions – select the best order of actions in order to reach a goal programmed into it – then the computer is doing a kind of reasoning. The computer is choosing which sequence of actions is best in order to reach the goal (productivity increases).

This process of determining the best order of actions to reach a goal is an essential feature of problem solving. Problem solving is getting closer to the higher level of reasoning called Generalization. But remember what we said above:

• To Generalize – to draw conclusions based on evidence and to test those conclusions through trial and error – requires intellectual flexibility and adaptation.

• Intelligence must be open (flexible) to new evidence and thus the need to revise (adapt) conclusions.

So let’s say a computer determines the best order of actions in mining, but in the process of performing those actions, something unexpected occurs. The environment in the mine suddenly changes, like water seeping into the mine or part of the mine becoming unstable. The unexpected change in the mine is “new evidence.” To do the higher-level reasoning called Generalization, the Artificial Intelligence of the computer would have to be open (flexible) to process this new evidence and to revise (adapt) its actions. This constant openness (flexibility) to new evidence and revision (adaptation) of conclusions adds extra layers of complexity and possibilities to the reasoning process. Let’s make the same point using a different example – the example of human language. We all know that a computer can respond in human language to a simple question like, “What is the weather in New York City today?” When the computer tells us the weather, it is not doing high- level reasoning. It is not really understanding human language in its complexities. Rather, it is simply performing a voice recognition of certain sounds – “weather,” “New York City,” “today”

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 6

– to which the computer is pre-programmed to respond with certain sounds. The same is true when a computer translates relatively simple sentences from one language to another. It is simply recognizing certain sounds and responding with a pre-programmed response. The level of AI in these cases is like Rote Learning. But let’s think through human language more carefully. The above question – “What is the weather in New York City today?” – is a simple question. Consider asking a more complex question like this: “How might a change in the pressure system to the north of the city affect the chance of the storm 200 miles south of the city from brining rain this afternoon?” Notice what the second question requires. It requires the intellectual flexibility to process new evidence – “pressure system to the north,” “storm 200 miles south,” – and then adapt (revise) the conclusion about today’s weather. Such flexibility and adaptation are characteristics of human intelligence – and thus features of human language. The more intelligence is flexible and able to adapt, the more it can engage complexity and consider unknown possibilities.

• Consider, for example, your own intelligence in engaging these lectures.

• Your intelligence is being flexible in processing new information and adapting (revising) your previous understanding. The more your intelligence is flexible and adapts, the more your intelligence engages complexity and considers previously unknown possibilities.

• Human language is a good example of the higher-level learning called Generalization because human language includes an almost unlimited variety of sentences and meanings, requiring constant intellectual flexibility and adaptation, and thus constant engagement with complexity and unknown possibilities.

It’s unclear how much Artificial Intelligence can mimic the intellectual flexibility and adaptation of human intelligence. AI has certainly come a long way in recent years. As our examples above indicate, AI plays important roles in manufacturing, mining, and language recognition. AI is also increasingly playing important roles in transportation (driverless vehicles) and financial transactions (buying and selling stocks). But as these examples also indicate, it’s unclear how much Artificial Intelligence can really resemble human intelligence by engaging in the higher level of reasoning called Generalization. What we can say is that the further development of AI will most likely be linked to the development of what’s called Quantum Computing.

Quantum Computing (QC) Quantum computing is still in the development stage, so its future is not clear. However, if it develops into a widely usable technology, it could represent a breakthrough in computer technology. Such a breakthrough would expand the possibilities of Artificial Intelligence. What is quantum computing? Let’s begin addressing this question by imaging a familiar scenario. Imagine flipping a coin in the air. Think of the coin as it’s flipping. As it flips, its result –

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 7

heads or tails – is not yet known. Its result is not yet determined, or indeterminate. This is called a “superposition” state.

• A “position” state means a clear result – heads or tails.

• A “superposition” state is not known, or not determined – indeterminate.

• The point here is simply to understand what “superposition” means. A flipping coin is in a superposition or indeterminate state.

Now let’s understand what the word “quantum” means. Quantum refers to something very small. It is associated with quantum physics. Quantum physics means studying the motion of very small particles like atoms – or even smaller (subatomic) particles like electrons and photons. In contrast, astrophysics studies the motion of big things, like planets, stars, and galaxies. One of the discoveries of quantum physics is that small particles like photons exist in superposition states. That means a photon is both a particle and a wave at the same time. A photon is simultaneously a particle of light and a wave of light. Its state is indeterminate. It is not only a particle or only a wave. It is both. And – this is key – it is both at the same time. That doesn’t seem logical. It’s not the way the physics of the visible world works. But it is the way the physics of the quantum world works – the very small world of subatomic particles. Subatomic particles exist in superposition – or indeterminate – states. Quantum computing is based on this strange nature of subatomic reality – the superposition states of subatomic particles. Let’s seek to understand this as simply as possible. Let’s first think about how a regular computer works. To do this, imagine you are in a maze – like a corn maze or a hedge maze. The walls are much taller than you. You’ll try to get out by choosing a path to walk down. If you reach a dead end, you’ll turn back and try another path. The point here is that you can only try one path at a time. You can’t walk down two different paths at the same time. This is how a regular computer works. It solves problems by trying one possible solution at a time. Regular computers run on integrated circuits of transistors. Transistors transmit electronic signals which encode information in “bits.” Each bit is represented as a number – either “0” or “1.” So one bit can be in two possible states – the “0” state or the “1” state.

• Imagine two bits in your computer. Two bits can be in four possible states – 00, 01, 10, or 11.

• But the two bits can only be in one of these states at a time.

• If a computer tries to solve a problem, it might put both bits in the zero state – 00 (like walking down one path of the maze). If that doesn’t solve the problem, it might then try the “01” state (a different path), then the “10” state and so on.

• In reality, regular computers process billions of bits – 0000101100111000………. But each bit can only be in one state at a time – either a “0” or “1.”

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 8

• Because each bit can only be in one state at a time, the computer can only process one sequence of bits at a time. Even if the computer is really fast – say it processes each sequence of bits in one-billionth of a second – it’s still processing one sequence at a time. That’s like you having super human speed in the maze. No matter how fast you are, you can only try one path at a time.

However, quantum computing is different. Quantum computing relies on quantum bits called qbits (or qubits). Qbits are part of the quantum world. Like photons, qubits can be in two separate states at the same time – a superposition state.

• Qbits can thus be both in the “0” state and the “1” state at the same time.

• Moreover, the superposition states of qbits are linked together in a process called quantum entanglement. This means that qbits can interact simultaneously with each other.

• The result is that quantum computers can process many sequences of qbits at the same time – not just one sequence with trillions of qbits, but process trillions of different sequences each of which contains trillions of qbits, all at the same time. (Like being able to try all the paths in the maze simultaneously.)

This means that quantum computers have the potential to be exponentially faster and more powerful than regular computers. To envision this increased speed and power, imagine the following. Imagine putting grains of wheat on the squares of a checkerboard.

• Put two grains of wheat on the first square. Then four grains on the next square, six on the next square, eight on the next, and so on, adding two extra grains to each new square.

o The total number of grains of wheat on the whole board would be 4,160.

• Repeat the process, but using exponents instead of addition. 2 grains on the first square, 2² on the next square, 2³ on the next, 2⁴ on the next, 2⁵ on the next, and so on.

o The number of grains of wheat on just the last square of the board would be 18,446,744,073,709,551,616.

o This number is exponentially bigger than 4,160. This kind of difference highlights how quantum computing has the potential, if successfully developed, to be exponentially faster and more powerful than regular computers.

The possibilities are extraordinary. The exponential growth in computer speed and power could take AI to a whole new level. As we said above, even though AI can already do impressive things, it’s unclear how much AI can really resemble human intelligence by engaging in the higher level of reasoning called Generalization. But if quantum computing is successfully developed into a widely usable technology, it could be a major step in the further development of AI toward higher level reasoning. Think through what this means. The successful development of quantum computing leading to major advances in AI • could mean major increases in productivity in all kinds of industries – manufacturing,

mining, health care, transportation, finance, etc.

• In other words, this could be the next major step in the history of industrialization, with technology increasing productivity, which increases wealth.

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 9

• Such increases in productivity and wealth (economic growth) would mean a higher standard of living.

Increasing productivity, wealth, and the standard of living would represent economic progress. Keep in mind, though, what we’ve discussed in earlier lectures – economic progress involves creative destruction and is thus multidimensional. It can also involve energy transitions. So the discussion of AI and quantum computing returns us to themes we’ve examined before – creative destruction and energy transitions. We will consider those themes again in the remainder of this lecture. Before we do so, however, we should note the following. As much as the “fourth industrial revolution” has the potential to bring significant economic progress, it also has the potential to bring significant dangers. Those dangers include increasing surveillance and data collection. Modern states and large corporations already engage in surveillance. They surveil our communications, consumption habits, and online patterns. For example, when you use platforms like Google, Facebook, Twitter, or YouTube, you as the user are the product being bought and sold. Your online behavior is collected as data and sold to third parties which use the data for commercial and political purposes. Advances in AI and quantum computing will vastly increase such surveillance power through technologies like biometric surveillance – tracking our unique physical and behavioral characteristics. The resulting data can then be used to reward or punish individuals for their behaviors and opinions. The Chinese Communist Party (CCP) is already doing this in China. The CCP has instituted a “social credit” system, making it easier or harder for individuals to access banks, credit, and educational institutions based on whether the individual’s behavior is acceptable to the Party. In other words, the CCP uses technology to punish individuals with “undesirable” opinions, firing them from jobs and banning them from college, essentially excluding them from regular society. What’s particularly troubling is that Big Tech companies work with the CCP in order to do this kind of surveillance in China. Big Tech companies thus have the ability to perform similar kinds of surveillance in the United States. This power to surveil – and to punish “undesirable” opinions – highlights the “dual nature” of technology. “Dual nature” means technology often has dual consequences. Technology can lead to positive things like increased productivity and wealth while it also leads to dangerous things like increased surveillance and control of human behavior. We will again address this issue of the “dual nature” of technology in a later lecture. For now, though, let’s complete this lecture by returning to central themes in the history of industrialization – creative destruction and energy transitions.

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 10

Creative Destruction (again) & Energy Transitions (again)

We’ve talked about creative destruction several times before. Our discussions have highlighted the following points:

• Creative destruction is necessary for economic progress. It is not possible to have the “creative” part – new technologies which help the economy grow – without the “destruction” part – destroying some existing jobs and businesses. An economy which is not destroying jobs is an economy which is not growing, and thus not improving the production of goods and services like education, food, health care, etc.

• The economic progress of creative destruction is multi-dimensional.

o In the long term – across generations – creative destruction benefits entire societies by raising the standard of living. An example is that the average standard of living of the poor in a developed society is higher than the average standard of living of the majority in an underdeveloped society.

o In the short term – within one generation – the effects of creative destruction are more complex. Many benefit in the short term. Inventors and investors in new technology make money. Consumers benefit by new products. And people with the skills to take advantage of the newly created jobs benefit. But others experience the destruction side of creative destruction. These include people who work at or own businesses which are destroyed, those without the skills to compete in a more advanced economy, and investors in previous technologies.

From the last bullet point, let’s highlight the difference between those with and those without the skills necessary to take advantage of the newly created jobs. Creative destruction creates new good-paying jobs which require specific skills as well as lower-paying jobs which require fewer skills. The higher paying jobs are also usually more interesting. With this in mind, let’s return to a point we noted near the end of last lecture.

• We said that the creative destruction which began with computer technology in the second half of the 20th century (Post-Industrial Society) was distinct from earlier rounds of creative destruction (Early and Later Industrialization).

• The difference was that the new jobs related to computer technology require more education and training than the industrial jobs in earlier eras. This increased education and training meant increased pay. The increased pay meant increased social inequality.

• In other words, the creative destruction of Post-Industrial Society led to greater disparities of income than earlier rounds of creative destruction. Individuals with the training and education to take advantage of the newly created jobs made (and continue to make) significantly more money than individuals with less training and education.

Developments in Artificial Intelligence and quantum computing will likely further contribute to this difference in income between highly skilled and lesser skilled labor. Again, those with the

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 11

training and education to take advantage of the new jobs of the “fourth industrial revolution” will make a lot more money than those with less training and education. This means that economic progress – new technology, advances in the production of goods and services, and higher standards of living – will probably lead to greater inequalities of income.

• To repeat the point: Economic progress will bring many highly paid and interesting jobs in the 21st century. These high paying jobs will require much more education or training than the industrial jobs of earlier generations. And these higher education and training requirements will likely result in more social inequality – i.e., many who complete the education and training will likely make significantly more money than many who do not.

o More social inequality does not necessarily mean that the standard of living for those with less education and training will go down. Their standard of living will likely go up. But it will go up less – in some cases much less – compared to those with more education and training.

• What’s unclear is whether further advances in computer technology will lead to new ways of obtaining the training and skills needed to compete in the 21st century. Throughout the second half of the 20th century, developing the skills required for the new jobs of Post- Industrial Society often meant earning degrees at traditional colleges and universities. The extent to which the “fourth industrial revolution” will create alternative ways to obtain training and skill is an open question. As technologies to create and deliver knowledge advance, different ways to obtain knowledge will likely multiply.

Other important questions going forward include the production of energy. As we’ve seen, economic progress throughout the history of industrialization has often been related to energy transitions. We’ve already described what an energy transition is and discussed examples of energy transitions in earlier lectures. Let’s now return to this issue to finish this lecture and our discussion of industrialization. We noted in an earlier lecture that the United State has been undergoing an important energy transition for a couple decades. That transition is the transition from oil to natural gas. The increasing use of natural gas is a major reason our CO₂ emissions have decreased since the start of the 21st century. Consider the following facts:

• CO₂ emissions from burning coal in the United States decreased by well over 50% in the first two decades of the 21st century. Emissions from coal decreased by 15% in 2019 alone.

• From 2016-19, the share of electricity generated by burning natural gas increased from 33% to 38%. The amount of electricity generated by burning coal decreased from 30% to 23%.

• These numbers show how many energy companies burn more natural gas, and less coal and oil, to produce electricity. Because of technological advances like hydraulic fracturing and horizontal drilling, natural gas is cheaper than coal and oil.

• Think about that last point. Price is key. As we’ve said in discussing earlier energy transitions, the use of technology to develop more efficient (cheaper) energy is important

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 12

to help the economy grow. Because technological advances have made natural gas cheaper, the increasing use of natural gas has put many coal plants out of business.

There are, of course, other sources of energy. Since the turn of the 21st century, the United States has increased the amount of electricity generated through non-carbon sources like nuclear and hydropower. Ironically, these sources – nuclear and hydropower – usually receive less attention than other potential sources labeled “clean,” “renewable,” or “green” energy. This is ironic since nuclear and hydropower are themselves renewable and clean. In fact, nuclear power already provides over half of our country’s “clean” (non-carbon) energy. And we are beginning to see incredible developments in nuclear energy such as miniaturized fission plants which are much smaller, safer, and cleaner than 20th century nuclear power plants. Still, though, media and political attention tend to focus on other potential sources of energy that are labeled “clean,” “renewable,” or “green.” One example is solar energy, which involves the use of solar panels. Solar panels use photovoltaic (PV) cells to convert sunlight to electricity. But many questions remain. Here are a few: • Why are solar panels not the primary source of energy for every building at Bronx

Community College, or every building in the entire CUNY system, or every building throughout NYC? The answer is efficiency and cost. Solar panels are not efficient enough at this point, meaning their cost is high in relation to their energy production. Maybe their efficiency will dramatically improve in the future.

• Another way of thinking this through is to ask the following question: Are solar panels the primary source of energy powering the factories which make solar panels? If not, why not? The answer again involves efficiency and cost.

Let’s dig deeper by thinking this through more. Discussions of “green,” “renewable,” and “clean” energy suggest that such energy does not depend on fossil fuels. But is this true? Is “green” energy really independent of fossil fuels? Let’s consider solar panels and wind turbines. Solar panels and wind turbines are made from minerals and metals mined from the earth. Examples of these minerals and metals are aluminum, cadmium, copper, gallium, indium, iron, lead, nickel, silica, silver, tellurium, tin, zinc, neodymium, terbium, indium, dysprosium, and praseodymium. Knowing that these minerals and metals are needed to make solar panels and wind turbines, here are some questions to consider in thinking through the relationship between “green” energy and fossil fuels:

• Do fossil fuels power the machinery used in mining the minerals and metals needed to make solar panels and wind turbines? If so, then “green” energy depends on fossil fuels.

• Do fossil fuels power the factories which make the machinery which mine the minerals? If so, then “green” energy depends on fossil fuels.

• Does the manufacturing of solar panels and wind turbines depend on the burning of fossil fuels? If so, then “green” energy depends on fossil fuels.

• Though solar panels and wind turbines are often referred to as “renewable” sources of energy, are the panels and turbines made out of non-renewable resources? If so – if the

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 13

panels and turbines are made from non-renewable resources – then are they really “renewable” sources of energy?

There are a couple reasons to ask these questions. One is to think through the relationship between “green/renewable” energy and fossil fuels. Another is to use what we’ve learned about energy transitions in the history of industrialization to help us think through current energy issues. After all, a good reason to study history is to use historical knowledge to help in thinking through current issues. So we’ve learned about several energy transitions in studying the history of industrialization.

• We’ve learned about the transition from wood to coal in early industrialization.

• We’ve learned about the transition from coal to oil in later industrialization.

• We also know that the United States has been undergoing a new energy transition from oil to natural gas since the beginning of the 21st century.

• These are all fossil fuels – coal, oil, and natural gas (though burning natural gas emits much less CO₂ than burning coal and oil).

Let’s now ask: Are we also going through a different kind of energy transition, a transition to non-fossil fuel sources like solar? The answer to this question is unclear at this point. The answer is unclear because of the information we’ve just discussed: 1) the making of solar panels remains linked to the burning of fossil fuels and 2) solar panels are not efficient enough at this point – their cost is high in relation to their energy production. The inefficiency of solar panels is the reason they are not used as the primary source of energy for businesses, industries, and other institutions throughout the country. Perhaps “green” energy will develop to be more efficient, more widely usable, and more independent of fossil fuels in the coming decades and generations. It’s just not clear at this point. Nuclear energy shows significant promise, though as noted it receives a lot less attention than other potential sources like solar. What is clear is that human beings face some hard choices about energy use in the short term. Consider underdeveloped parts of the world, such as many countries in Africa. Millions of Africans die every year because their societies lack enough energy to fuel economic growth. Much of Africa lacks reliable sources of electricity. As a result, hundreds of millions of Africans still burn solid fuel — wood, dried dung, and charcoal — for heat, cooking, and light. These solid fuels are often burned inside, causing massive health problems which result in millions of premature deaths. Africa has vast amounts of coal reserves. It requires a lot of money to develop these coal reserves and build coal-fired power plants, which could generate electricity and help lift millions of Africans out of grinding poverty. But the World Bank and many Western governments lend money only for the production of “green” energy. They restrict lending

This content is protected, and may not be shared, uploaded, or distributed. Fourth Revolution - 14

money to maintain or build new coal-fired power plants. As a result, many African nations are turning to China to finance the creation of coal-powered energy. The focus on “green” energy means that hundreds of millions of Africans will continue to burn solid fuels. They have little choice, even though burning solid fuels contributes to a range of diseases like pneumonia, heart disease, pulmonary disease, stroke, and lung cancer. What this discussion of Africa highlights is that energy use, like economic progress, is multidimensional. Energy issues – fossil fuels, “green” energy, etc. – are not one-dimensional issues with easy answers. Yet the efficient production and use of energy has been and remains crucial to lifting human beings out of grinding poverty. Remember, poverty is the normal condition of humans throughout human history. That only began to change as the more efficient use of energy – starting in early industrialization – began to create wealth. We have inherited this wealth-creating process. We are part of a tiny fraction of all humans in human history who do not simply accept poverty as the inevitable condition of human beings.