George Gilder's brilliant annual tech summit COSM Summit gives us a glimpse of a future that excites many, but also scares some. It is an optimistic view of the future that is often very different from that presented by short-sighted forecasters. However, one thing that will not change in the future , regardless of the technological outlook, is the central role that energy plays in supporting it.
Energy physics brings it all together. This also applies to the theme of this year's COSM Summit; The manifestation of AI in nature, the potential of graphene as a truly new and revolutionary material, and China's role on the world stage.
This may be obvious, but it goes without saying that not only society, but life as we know it, and certainly the universe, would not exist if only energy existed. Not to get too philosophical, but there's no way around it in Gilder's COSM, all possible futures occur at the intersection of three primary properties of reality: information, atoms, and energy. As George said earlier, the only thing that separates us from Neanderthals is what we know . The building blocks that make up everything that we all exist at that time. Today we have more information about "just" atoms and the same forces that have always existed.
We have the ability to increase the amount of information about the rearrangement of nature's atoms in unique and even miraculous ways that will enable humanity to create all current and future products and services. But receiving and processing this information, as well as using this information to rearrange atoms, always requires energy. Energy is consumed by every innovation, product and derivative service that makes life interesting, safe, comfortable, enjoyable and even beautiful.
And throughout history, inventors have found more ways to make things that consume more energy than they produce . The discovery of materials such as hybrids, polymers, drugs or monocrystalline silicon has led to the need for new energy for their production. Likewise, the invention of machines made from these materials, such as cars, airplanes or computers, has created new energy demands.
The inevitability of natural energy is manifested in the fact that our own information machinery has become extremely energy-intensive. Any software, even virtual reality, needs a power-hungry reality machine to generate logic. This may seem obvious, but the implications of this fact mean that the global cloud today, roughly speaking, consumes as much energy as global aviation. Additionally, the former growth is much faster than the latter growth.
This brings us to the focus of this meeting on AI, a whole new way of driving silicon machines. Although AI has been around for a while, November 30, 2022, when ChatGPT became available to the public, will be remembered as the date AI entered the wild.
AI is the most powerful silicon application ever created. In terms of power, it is similar to the jet aircraft of the age of steamships. The latter encourages personal travel around the world because it is much better, more comfortable and therefore more efficient, which means it saves the most precious resource in the universe: our time. Of course, flying is a more energy-intensive way to transport things. AI does the same.
AI naturally consists of a training phase, or what is called a machine learning phase, and an inference phase, which involves the application of acquired knowledge. Both education and distraction consume a lot of energy, at least for those obsessed with reducing society's energy consumption. For example, a small machine learning algorithm trained a few years ago to solve a single puzzle—a Rubik's Cube—spent enough electricity to drive a Tesla millions of miles. And learning is not a one-time event for many real-world problems. Then, after exercise, comes the final phase, which, despite being less energy intensive than exercise, is performed more frequently, sometimes continuously, and results in a greater overall energy expenditure than exercise. Think of it as comparing the energy needed to make aluminum and build an airplane and then the fuel needed to fly it.
The full impact of artificial intelligence in energy conservation is yet to be seen as software and hardware development is still in its infancy. We live in the age of conventional calculators, say around 1980. Dystopians fear artificial intelligence, but it is an exciting and important new tool that will unlock all kinds of innovation, not just self-driving cars and robots, but other innovations that will become us. And greater productivity, and new ways of identifying keys, as well as many things you never thought of. And the infrastructure that will be built and strengthened to democratize and enhance the usability of AI will use the Internet, to borrow a term from Andreessen Horowitz.
At a recent meeting of electric utility executives, Elon Musk mildly criticized them for underestimating the level of future electricity demand. He wasn't talking about electricity, but mostly about AI power. By comparison, the global cloud currently consumes 10 times more electricity than all the world's electric cars combined. Even if electric vehicle adoption grows at the same pace as investors believe, cloud technology will still outpace electricity demand, especially as AI hardware is rapidly being incorporated into cloud infrastructure.
The standard response to observations about the power intensity of computing, and artificial intelligence in particular, is that innovators will make silicon technology more efficient. Of course they will. But instead of slowing the growth in energy demand, efficiency is actually driving that growth. This fact is called Jevon's paradox. Information systems in general are the most striking example of this paradox.
Consider that the power efficiency of logic motors has increased more than a billion times over the past 60 years. And that's why there are billions of smartphones and thousands of warehouse-scale data centers today. By the 1980s, computing power efficiency allowed a single smartphone to use more electricity than the buildings we live in today, and today a single data center would require the entire American power grid. In other words, the era of smartphones and the cloud would not have existed without the phenomenal growth of computing power.
Something truly extraordinary is happening now, discovered by chance two decades ago. An entirely new and game-changing class is called graphene, a material made from nearly infinite sheets of pure carbon, just a few layers thick, with seemingly magical properties. Graphene is beginning to be used in various commercial products. One possibility is to use graphene as a much more efficient base material to replace silicon in computer chips. Rely on Jevons' paradoxical acceleration.
Graphene has a series of other obvious structures and properties relevant to biology. On the one hand, it is much stronger than steel. In other formulations, the drug is biocompatible and can promote nerve recovery that was previously impossible. And graphene is one of many new classes of materials emerging from laboratories, though it may be the most unusual.
But let's get back to our topic. All materials require energy to create. Compared to a few centuries before the modern period, almost all were made of some materials, mostly stone, wood and animal parts. On average, more than one kilogram of energy is required to make the materials of our time. If we switch from wood to polymers, which are widely used in medicine and are much more useful than wood, the energy value per kilogram produced will increase by 10 times. Use aluminum instead of polymer and the energy value per kilogram increases by 10 times. And semiconductor grade silicon has a power density 30 times higher than aluminum. Producing one kilogram of silicon requires 100 times more energy than producing one kilogram of steel. And the world produces kilotons of silicon, the energy equivalent of producing megatons of steel, not only for computer chips but also for solar panels.
When it comes to producing graphene, we are just beginning to understand how to produce it on a large scale. George Gilder suggests that graphene production may soon occur at the "aluminum moment," a point in history dating back to 1886 when inventors found a way to make an attractive but very expensive material at an affordable price. Earlier, pure aluminum was more expensive than gold.
However, as the technical literature shows, the forming strength of graphene is more similar to that of silicon than that of aluminum. Therefore, I argue that graphene is not at the "aluminum moment" but at the Czokralski edge. Polish metallurgist Jan Zocralski's moment In 1916, he accidentally discovered how to make monocrystalline silicon from a melt. This discovery led directly to the commercial process today, which was perfected by Bell Labs in 1949, 33 years after the accidental discovery of silicon. Without monocrystalline silicon, there would be no era of silicon computers. If graphene takes this long to go from a opportunistic discovery to a viable commercial process, we'll have to wait another ten years. But now, in the delicate cycle of materials, machines and information, the AI-powered supercomputers of our time will likely help shorten that time radically.
The first companies and countries to achieve commercially viable graphene at scale will have real benefits. This brings us to China's energy modernization, the third of COSM 2023's three themes.
Consider the situation with several basic materials that require a lot of energy to produce and, as a guarantee, are essential for energy-producing and energy-consuming vehicles.
China produces more than 60% of the world's aluminum, processes more than half of the world's copper, a material that forms the basis of 90% of electrical products, and 90% of the world's refined rare earth elements, an essential element for many people to generate electricity. engine or generator. And essential in many high-tech applications, including solar panels and wind turbines, 90% of Earth's gallium has been released, making possible the magical gallium arsenide semiconductor, which is used to make many technology products, especially lasers and LEDs. . and 60% of the world's refined lithium, 80% of the world's refined graphite used in all lithium batteries, and 50% to 90% of the most basic chemical and polymer components needed to make lithium batteries. Not only that; But you get the point.
China is not afraid of industries that produce energy-intensive materials and two decades ago wanted to become a dominant supplier of such materials. This leadership is at the intersection of three principles that encourage engineers to study basic chemistry, electricity, and ancient materials, which are second or third priorities here, and politics comes second. It facilitates and accelerates industry's ability to build chemical and energy-intensive facilities, rather than resisting and hindering them as we do in the United States, and third, policies that ensure a reliable supply of cheap energy. This feeds into industrial facilities. In China the latter means a grill that is two-thirds charcoal powered.
Now the United States is preparing the Deflation Act, the largest industrial policy spending package in US history. It's no secret what the specific purpose of most IRA spending is; It helps reduce the country's carbon emissions by promoting energy transition from hydrocarbons. Whatever you know about climate change and carbon dioxide, there are two things to consider in the relationship between technology, policy and energy.
The first incident was revealed.
The $2 trillion still to be distributed by the IRA is expected to reduce US CO2 emissions by 1 gigaton per year if government estimates prove correct. Compared to the theoretical emission yield and from a practical point of view, China is still building more coal-driven power plants, and President Shi has made it clear that they will do so. This means that China will continue to enjoy an industrial energy consumption benefit for the power-intensive industry over the next decades. This means that the completion of the extra coal-powered power plant will add 2 more gigatons every year to the country's already high CO2 emission. Meanwhile, if a theoretically excludes 1 gigaton here, the majority of US taxpayers will be used to buy important energy materials from China for the construction of air, solar and battery infrastructure for the IRA for the IRA purpose.
And the second event is like remembering.
Keeping China and its fuel materials at the center of energy transformation, global expenditure on air and solar energy and at least two decades after two decades of hydrocarbons and coal lifting, at least $ 5 trillion, is the world's condition today. Oil, natural gas.
These costs reduce the energy portion produced by hydrocarbon, but by only two percent points. Today, hydrocarbons still provide 82% of the world's energy. And combined investment to solar and air power equipment currently provides less than 4% of global energy. In comparison, wood burning still provides 10% energy in the world. Meanwhile, in the last two decades, the absolute amount of hydrocarbon in the world in the form of energy has increased, actually increased by Saudi Arabia's oil production by six times .
The evidence is clear that spending on "Change of Power" has created bad results so far. It is clear in our policy that we can spend more money to launch carbon-free vehicles. But none of the physics of the powerful substance or the main supplier of such materials can be ignored.
As cosm 2023 explains, extraordinary transformative innovations are emerging in both computing and materials. But these revolutions have come up again innovation that uses instead of energy generation.
Energy- The nature of the producing machines and the more changes on the scale awaits the unknown future success, revelation or accidental discovery. Such progress is inevitable, but according to Bill Gates, such a revolutionary progress is "no prediction work."
But it can be predicted that artificial intelligence infrastructure will spread rapidly and one day someone will find a way to make graphin in a large range. In terms of geopolitical situations, we can expect, perhaps hopefully, the wisdom of returning to the state of art policy.
