November 2031, Rocket City, Bright Country
Michael and Daphne hosted Irene at their home. As soon as they met, Daphne asked with concern, "Andek is seriously ill. Why did you come back to Bright Country? Is he feeling any better?"
"His condition was quite serious for a while, but he's improved and stabilized now. I came back to Bright Country because a research project here needs to be wrapped up," Irene said. "Andek is also concerned about the letter he sent Michael. He asked me to hear what Michael thinks."
Michael looked a little awkward. Is Irene here to deliver a final ultimatum on behalf of the media mogul?
He said, "I'm very grateful for Andek's concern. His letter moved me deeply. I know he wasn't doing it for himself."
Irene looked up at Michael, waiting for him to continue. Daphne smiled and stepped in to break the awkward silence. "I know Michael's answer. Guess what? It actually has to do with another letter you gave him."
"Another letter? You mean Mirov's Maya Notes?" Irene asked.
Daphne enthusiastically told Irene about the contents of the Maya Notes. Captain Mirov had studied the Maya Calendar and its corresponding planetary synodic periods in depth, and reached a startling conclusion.
Besides the base-twenty system they used in everyday life, the Maya had a particular fondness for special values expressed in ternary, whose symmetry they found beautiful. For example, 13=(111)3; the total number of pyramid steps was 364=(111111)3; and the least common multiple of the orbital periods of Earth and the synodic periods of Mercury, Venus, Mars, and Saturn relative to Earth was 7174440=(111111111111000)3.
Venus's synodic period with Earth was 584 days. The greatest common divisor of this number and 365 was 73, and 584/73=8. The Maya thought this coincidence was remarkable. They regarded 8 as Venus's contribution number and recorded it on stone tablets. In ternary, 8=(22)3=(1.0-1)3=(10T)3, which was symmetrical.
Inspired by the number 8, Mirov hypothesized that the Maya might have considered 8 a wondrous and special value because they had noticed that its ternary representation displayed a symmetry even more beautiful than the repeated symmetry of 3 ones, 6 ones, 12 ones, and so on.
Were two 2s logically more elegant than three 1s? Clearly not, because they were both simple repetitions and therefore identical in logical form.
So what made 8 special? In ternary, there were only three symbols: 1, 0, and -1, denoted by T. The number 8 used all three symbols, in a centrally symmetrical form: 8=(10T)3.
Any integer n could be expressed in any base. In general, a base was a positive integer, represented by b; the number of digits in n was represented by m. Then n=∑a*b^m, where a ranged from 0 to (b-1), and m ranged from 0 to m-1.
Since b-1 was congruent to -1 modulo b, the 9 in decimal could be defined as -1, while -1 in base twenty would be 19. When b=3, that is, in ternary, b-1=2, so 2 was defined as -1.
Ternary under this convention used only 1, 0, and -1, and was called symmetric (balanced) ternary. In computers, because -1 consisted of two characters, the letter "T" was conventionally used to represent it.
8=(22)3=(1.0-1)3=(10T)3. This showed that in ternary, 8 was both the most comprehensive and the most concise higher-order symmetrical form, containing all three symbols: 1, 0, and -1.
Irene was a little dazed by the explanation. It was so complicated. Weren't the 0 and 1 of binary simpler? Why introduce -1?
Michael suddenly asked out of nowhere, "Irene, we're having a get-together next Wednesday evening. Can you come?"
Caught off guard, Irene had no idea what he was getting at. She thought over her schedule. It was only Tuesday, and she hadn't planned out next week yet.
As she pondered, she replied, "I'd love to come to your get-together, but I'm not sure if I can."
Michael and Daphne exchanged a smile. He said to Irene, "My question and your answer are very common in everyday life. They stem from the way the human brain thinks. Ask it a question, and the answer isn't always just yes or no. In many cases, there's a third answer: uncertain."
Irene suddenly understood. "In symmetric ternary, 1 means yes, T means no, and 0 means uncertain. This way of making decisions is obviously better suited to the human brain. There are so many uncertain cases in my genetic engineering research, but I never thought of using ternary."
Daphne answered the question Irene had come to ask on Michael's behalf. "A man with a dream is always stubbornly persistent. At first, I was worried that Michael would say no when Andek urged him to give up the Mars terraforming plan. Luckily, his answer is uncertain for now."
Irene pressed on. "Whether I can come to the get-together next Wednesday is uncertain for now, but in a few days I'll give you a definite yes or no. You can't keep your decision about continuing or abandoning the Mars terraforming plan uncertain forever, can you?"
Michael's expression turned somewhat serious. "Turning Mars into a 'backup Earth' would be a tremendous achievement for humanity, with benefits too vast to measure. Of course, it would also involve risks. I plan to upgrade computers to ternary and run a simulation using their vastly greater computing power. Once I get the results, I'll be able to decide."
Irene remembered Andek's instructions before she left. "We only have one Earth. Andek warned you that even a one-in-a-million risk wouldn't justify building the so-called Superlight Wave Power Station. No matter how good a simulation is, it can't prove that the risk is zero, can it?"
Michael thought for a moment, then gave an example. "Humanity has witnessed the harm caused by the atomic bomb. But nuclear power is clean energy. Should we build nuclear power plants? The risks of nuclear power plants aren't just theoretical; nuclear disasters have happened in real life. Yet even today, there are still hundreds of nuclear power plants operating around the world."
The three fell silent. It seemed that persuading one another wouldn't be easy, so the conversation naturally shifted to ternary computers.
In the 1970s, the Former Soviet Union halted its research and development of ternary computers. They were short on money and lacked a complete industrial supply chain. Most importantly, Bright Country and Europe hadn't blocked the Former Soviet Union from accessing binary computer equipment and applications.
The Soviets realized that since they could buy what they wanted, the cost was far lower than investing heavily in building an entire ternary system from scratch. Why wouldn't they take the easier option?
Meanwhile, the computer industry based on binary developed rapidly. Transistors replaced vacuum tubes, the density of integrated circuits per unit area kept rising, and computing speeds increased exponentially. Incredibly, Moore's Law, measured in years or even half-years, continued for decades.
Storage, computation, transmission, and packaging technologies advanced rapidly, while new materials and processes emerged one after another. The internet, mobile internet, large AI models, AGI, and various smart devices such as intelligent robots took turns in the spotlight, driving ever-greater demand for computing power.
At last, Moore's Law neared its physical limits. Once integrated circuit widths had shrunk from tens of nanometers to just a few, existing processes could no longer support denser layouts.
Large AI models seemed to open their cavernous mouths and devour humanity's already insufficient electricity resources.
Low-energy, power-saving computing solutions became a priority, and ternary architecture once again became a hot research topic.
In theory, for the same computational power per unit area, integrated circuits using ternary architecture were less dense than those using binary, giving them a clear low-power advantage. Conversely, at the same integrated circuit density, ternary architecture computed faster than binary.
But ternary architecture had to be built from the ground up. It required enormous additional investment and faced many challenges. The first weak link in ternary architecture was the materials used for its components.
There were many competing technological approaches to ternary components, but they could be divided into two main categories.
One used carbon nanotubes. By controlling them at the nanoscale, tubes with different diameters could output three stable voltages—high, medium, and low—to represent 1, 0, and -1, respectively. This was called the "tube-diameter method."
The other approach stacked three different metals and oxides—for example, metallic lithium, lithium phosphate, and metallic nickel—each outputting a different voltage to represent one of the three states. This was called the "stacking method."
Both approaches could significantly reduce power consumption and increase computing speed. The conversions between the three voltages used in each approach were reversible and repeatable, which was essential for ternary computer components.
Ternary architecture had a clear low-power advantage and could save electricity. But large AI models consumed not only huge amounts of electricity, but also huge amounts of water.
In many countries and regions, water was even scarcer than electricity.
As integrated circuits became packed at the atomic scale, heat dissipation became a major problem during the production and operation of nanometer-scale chips.
Traditional fans could no longer meet the need, so ingenious engineers submerged entire circuit boards in a specially formulated "water" and used circulating water for cooling. This was known as "immersion cooling."
How could water consumption be reduced?
Amid the boom in electronic information technology, transistor technology continued to advance. The first-generation transistor was silicon, the second was gallium arsenide, and the third was silicon carbide.
New technological demands and new materials often advanced hand in hand, each giving rise to the other. Sometimes, though, the situation was awkward: a new material might be invented without any demand for its applications, while some applications had no suitable materials available.
For a long time, the computer industry had little demand for heat-resistant components.
This was partly a matter of cost and partly a question of whether they were necessary for any real-world applications. No one put computers over a fire, so what use were heat-resistant components?
For several decades after the middle of the twentieth century, the Former Soviet Union made more than twenty attempts to send probes to Venus, but almost none managed to transmit signals back successfully. The surface of Venus was hot and under high pressure, with temperatures exceeding 400 degrees. Any electrical equipment available on Earth would be destroyed by the heat and rendered useless.
New demands called for new materials. Silicon carbide was heat-resistant, had low electrical resistance, was hard, and was highly stable. Components made from silicon carbide could withstand temperatures of 500 degrees.
Almost all the electrical equipment on the Bright Country Space Administration's upcoming Venus probe had been switched to silicon carbide components.
Daphne knew a little about the new Venus probes. She said to Michael, "Your idea isn't just to use silicon carbide components in Venus probes. You also want to use them extensively in computer production and operation, so they can withstand high temperatures and reduce cooling needs, saving water."
Michael smiled mysteriously. "If the components can withstand high temperatures, we can switch from water cooling back to air cooling. That wouldn't just save water. We wouldn't need any water at all."
"Wow! How did you come up with that? It seems that innovation needs not only scientists, but also strategists who can set the direction!" Daphne exclaimed.
A poem assembled from lines at the end of the chapter:
No need to compare with plain silk, —Du Fu, Tang
Luckily, this stone surpasses it. —Xin Qiji, Song
Once, I had never set out, —Huang Shumei, Song
Knowing when to advance or withdraw is what matters. —Jiang Zaiheng, Qing
Before you continue