Energy Realities At The Nexus Of TechnoOptimism
By George Gilder The COSM Summit Each annual gathering of high-tech figures offers a glimpse into the future that many fear. This is a completely different optimistic view of the future offered by many short-sighted prognosticators. One thing that will not change in the future, regardless of the technological point of view, is the central and inspiring role of government.
The physics of energy ties everything together. This applies to the topics of the COSM meeting this year: the development of artificial intelligence "in the wild", the prospects of graphene as a completely new and revolutionary class of materials, as well as China's role on the world stage.
This may be obvious, but it should be said that not only society, but life as we know it, and the universe itself, would not exist without energy. I won't philosophize, but with Gilder's COSM it is impossible to avoid: all possible options for the future arise at the intersection of three fundamental properties of reality: information, atoms and energy. As George said earlier, the only thing that stands between our time and the Neanderthals is what we know . The building blocks that made up everything we had back then. We "simply" have more information about atoms and the same forces that exist today.
It is our ability to track the amount of information about how to rearrange nature's atoms in unique and wonderful ways that will enable humanity to create all current and future products and services. But obtaining and processing this information, as well as manipulating atoms using this information, always requires the use of energy. Energy is expended in every innovation, product and therefore service that makes life enjoyable, safe, comfortable, enjoyable and beautiful.
Throughout history, inventors have found more ways to create objects that consume energy than create energy. The discovery of materials such as composites, polymers, pharmaceuticals or monocrystalline silicon has led to the need to use new energy for their production. Similarly, the invention of machines made from these materials, such as cars, airplanes or computers, has created new demands for energy.
Our computing machines are very energy intensive and are an example of natural energy that cannot be ignored. Any software, even virtual reality, requires a reality machine that consumes electricity to accelerate the logic. This may seem obvious, but the bottom line is that today the global cloud is generally as powerful as global aviation. And the first develops faster than the second.
This brings us to the artificial intelligence at the heart of this meeting: a whole new way to control silicon motors. Although AI will be around for a while, November 30, 2022 will be the launch day for AI when ChatGPT is announced.
AI represents the most energy-intensive use of silicon in history. The transition from the age of steamships to jet aircraft was similar in power. The last world starts personal travel, because it is much better, more comfortable, and therefore more productive, that is, it saves us the most precious time in the universe. Of course, flying is a more energy-intensive way to move anything. The same goes for AI.
Artificial intelligence involves a training or machine learning stage, followed by an evaluation stage where the knowledge gained is used. At the very least, educating and raising awareness among people concerned about reducing society's energy consumption consumes a lot of energy. For example, a few years ago, a simple machine learning algorithm trained to solve a Rubik's Cube used enough electricity to drive a Tesla car a million miles. And for many real-life professions, learning doesn't happen once. Then, after exercise, comes the reference phase, which, although more energy-efficient than exercise, is performed very frequently, sometimes continuously, and results in a higher total energy expenditure than exercise. Think of it as comparing the energy needed to produce aluminum to the energy needed to build an airplane and then the fuel needed to fly.
As software and hardware development is still in its early stages, the full impact of AI on energy consumption remains to be seen. Around 1980, we were living through the era of mass computing. Anti-utopians fear artificial intelligence, but it is an exciting and sustainable new tool that will enable all kinds of innovation, not just autonomous cars and robots, but also increased productivity and new methods of fundamental discovery, and many things previously unimaginable. And the infrastructure that will be built and developed to democratize artificial intelligence will use networks adapted from Andreessen Horowitz.
At a recent meeting of power company executives, Elon Musk mildly criticized them for underestimating the amount of demand for electricity. It's not about the power of electric cars, but most importantly - about artificial intelligence. For example, the global cloud now consumes 10 times more electricity than all electric cars combined. Even if EV adoption grows rapidly, bullish predictions are that the cloud will still far outpace electricity demand, especially as AI hardware is rapidly being added to cloud infrastructure.
A common response to observations about the energy demands of computers and especially artificial intelligence is that innovation is making silicon technology more efficient. Of course they will. However, efficiency does not reduce the growth of energy demand, on the contrary, it encourages it. This fact was called the Jevons paradox. Information systems in general are the most interesting examples of paradoxes.
Consider that over the past 60 years, the energy efficiency of logic engines has grown more than a billion times . And that's why there are billions of smartphones and thousands of data centers the size of warehouses today. In the 1980s, the efficiency of computing power meant that smartphones used more electricity than existing buildings, and today a single data center requires the entire US electrical grid. In other words, without advances in incredibly energy-efficient computing, there would be no smartphone or cloud era.
And now an unusual object, accidentally discovered two decades ago, has appeared. A brand new and revolutionary material called graphene - an extremely thin layer of pure carbon only a few atomic layers thick - has seemingly magical properties. Graphene is beginning to be used in several commercial products. One option to replace silicon in computer chips is to use graphene as a high-performance base material. Trust us and we'll introduce you to the Jevons Paradox.
Graphene has many other conductive properties that are relevant to structure and biology. In some forms, it is stronger than steel. In other formulations, it can contribute to biochemical and pre-existing nerve regeneration disorders. And graphene is only one, although perhaps the most popular, of many new classes of materials produced by research laboratories.
But back to our topic: the production of all goods requires work. Compared to the era that preceded modern times, almost everything was built using fewer materials, such as stone, wood, and animal parts. Modern materials require more than an order of magnitude more energy to produce an average kilogram. The transition from wood to polymers, which are widely used in the medical sector and are more useful than wood, increases the value of each kilogram of energy produced by 10 times. If aluminum is used instead of polymer, the energy value of one kilogram of the product increases 10 times. And semiconductor silicon is 30 times stronger than aluminum. It takes 100 times more energy to produce one kilogram of silicon than to produce one kilogram of steel. And the world produces kilotons of silicon (in energy equivalent, it is equivalent to the production of one megaton of steel) not only for computer chips, but also for solar panels.
When we created graphene, we were still in the early stages of figuring out how to mass-produce it. George Gilder suggests that the production of graphene may have preceded the "aluminum moment". In 1886, when inventors reached the limit of producing an economical and attractive material in large quantities, it was attractive but very expensive. In the past, the price of pure aluminum was more expensive than gold.
However, according to the technical literature, the yield strength of graphene is more similar to that of silicon than that of aluminum. Therefore, I claim that graphene is not in the "aluminum moment", but in the Chachralsky peak. moment Polish metallurgist Jan Czachralski In 1916, he accidentally discovered how to obtain single crystal silicon from a molten bath. This discovery led directly to the commercial process used today, which was completed at Bell Labs in 1949, 33 years after silicon was accidentally discovered. Without monocrystalline silicon, there would be no era of silicon computers. If it had taken the same amount of time from its accidental discovery to the commercial processing of graphene, we would have had to wait ten years. But today, in the cycle of materials, machines and information, modern supercomputers of artificial intelligence help to shorten this period even more.
These companies and countries will have some real advantages in the large-scale commercialization of graphene in the first place. The third of the three themes of COSM 2023 brings us to the relevance of Chinese energy.
We will look at a number of key materials whose production requires high energy consumption and which are fundamental to machines that produce and use energy in parallel.
China produces more than 60 percent of the world's aluminum, processes more than half of the world's copper -- a critical element that provides nearly 90 percent of electricity -- and 90 percent of the world's refined rare earth metals, which are needed for many electric motors and generators. . Indispensable in many high-tech applications, including solar panels and wind turbines, 90% of the globe contains pure gallium, the magical gallium arsenide semiconductor used to make many technological objects, not least lasers and light-emitting diodes; And 60% of the world's purified lithium, 80% of the world's purified graphite is used in all lithium batteries, as well as 50% to 90% of many of the key chemical formulas and polymer components needed to make lithium batteries. That's not all; But you get the point.
China is not afraid of energy-intensive materials industries and twenty years ago decided to become their main supplier. This leadership arises from a combination of three types of policies: policies that encourage and encourage engineers to study the basic sciences of chemistry, electricity, and materials, which are at best second or third priorities, and, second, policies that encourage and accelerate— rather than resisting and obstructing as we do in America. - The ability of industry to build large chemical and energy facilities and, thirdly, a policy that guarantees a reliable supply of low-cost energy for the operation of industrial enterprises. The latter in China means a two-thirds coal-fired power system.
The United States has now enacted the Deflation Act, the largest industrial policy spending package in American history. It is no secret that much of the IRA's spending has a specific purpose: it serves to reduce the country's carbon emissions by stimulating the energy transition away from the use of hydrocarbons. Regardless of what you know about climate change and carbon dioxide, two facts about the intersection of technology, politics, and energy are worth considering.
The first group of facts:
An estimated $2 trillion in IRA investments would reduce US carbon emissions by 1 gigaton if the government's projections are correct. Leaving aside the effects of inflation, which is tied to theoretical emissions and in practice, China is still building more coal-fired power plants, and President Xi has made it clear that they will build more coal-fired power plants. This means that China will continue to benefit from industrial energy spending in energy-intensive industries for decades to come. This means that the completion of additional coal-fired power plants will increase carbon emissions, which are already the highest in the world, by 2 gigatonnes per year. Meanwhile, in theory, to remove 1 gigaton here, $2 trillion of US taxpayer funds would be spent on purchasing critical energy materials from China to build the wind, solar and battery equipment that the IRA targets.
And the second fact to remember:
As China and energy materials inevitably take center stage in the energy transition, it's worth noting where the world is today, two decades from now, with $5 trillion in global spending on wind and solar, as well as hydrocarbons, coal and oil. , natural gas. .
These costs reduce the share of energy produced by hydrocarbons, but only by a few percentage points. Now, hydrocarbons still provide 82 percent of the world's energy. And the combined contribution of solar and wind energy now provides less than 4 percent of the world's energy. For example, burning wood still provides 10 percent of the world's energy. Meanwhile, over the past twenty years, the absolute amount of hydrocarbons used in the world has grown not in percentage terms, but in energy terms, which is equivalent to six times the amount of oil produced by Saudi Arabia.
The data shows that spending on "energy transfer" has so far produced poor results. That we can afford to spend more money on the implementation of non-hydrocarbon energy engines, and perhaps this will be reflected in our policies. However, it cannot be denied that neither the physics of energy materials nor China are the main suppliers of these materials.
As explained at COSM 2023, both computing and materials are seeing truly amazing and game-changing innovations. However, these revolutions were again innovations that absorbed increasing power, not electricity production.
Changes in the nature and scale of energy-generating machines await breakthrough, enlightenment, or breakthrough into an unknown future. Such growth is inevitable, but in the words of Bill Gates, such revolutionary growth is "unpredictable".
But we can predict with confidence how the infrastructure of artificial intelligence will develop and how graphene will one day be made on a large scale. We can hope - perhaps even more - that common sense will return to the realm of industrial policy, given the geopolitical situation.
