362: Chapter 361 Three-Dimensional Transistors

To make Ling fully simulate hundreds, or even thousands of different personality traits, the required computing power would be immense.

According to Ling's preliminary estimates, this amount of computation would even exceed the combined computing power of the world's top three supercomputers.

And that is assuming the requirements do not continue to increase later on.

If Wang Shuo wanted Ling to match external traits like logic, emotions, actions, language, and so on, according to the characteristic encoding after randomly generating personality traits, the required computing power would increase exponentially. By then, let alone one data center, even giving Wang Shuo ten data centers probably wouldn't be enough.

Data centers are not only expensive to build and have long construction cycles, but even after they are built, the fragile equipment requires high maintenance costs. To say nothing else, just a set of temperature control equipment alone would deter ninety percent of companies.

Without an investment of at least ten billion, it's simply impossible to get it off the ground.

And the most frustrating part is that all servers are major energy consumers. If Wang Shuo were only building a standard enterprise-level server cluster, that would be one thing, but if he were building a super-large data center, the energy consumed just to keep these servers running would be an astronomical figure.

Therefore, most countries simply cannot afford this thing.

There is no special reason; it is simply a lack of electricity.

Building a data center within China is not a big problem; energy is sufficient, technology is not lacking, and industrial support is available. Relying on the relationship with the cheap senior fellow apprentice Wang Shuo just met, he might even be able to secure extra land subsidies and policy incentives.

All in all, there are only two factors restricting Wang Shuo from building a data center.

The first is cost.

Wang Shuo honestly cannot afford, nor does he want to pay, the cost of hundreds of billions.

Spending hundreds of billions just to build one data center is practically a crime, isn't it? It's a massive waste of resources!

The second is technology.

Don't misunderstand; it's not that Wang Shuo lacks the technology, but rather that he has no desire to use existing server technology to build a data center.

The equipment is bulky, the supporting facilities are complex, the footprint is huge, and construction and maintenance costs are extremely high. The key point is: the energy consumption is super high, yet the performance is extremely limited.

It can be said that in Wang Shuo's eyes, this thing has almost no advantages, but many disadvantages. If he had no options, that would be one thing, but now that the cards in his hand are about to be complete, who would still hold their nose and use this thing?

Wouldn't it be better to get it right in one step?

Thus, the Super Chip 2 was born.

In a sense, this Chip is the true Super Chip; it is what Wang Shuo had envisioned since he started his business but could not implement.

Its biggest difference from the Super Chip 1 is that Wang Shuo completely changed the underlying operation method of traditional Chips.

When manufacturing the Super Chip 1, because the manufacturing equipment was heavily modified and many aspects were not up to the original design, plus Wang Shuo wanted to match the existing Chip system for better testing, he did not rashly adopt the manufacturing route of three-dimensional transistors. He basically followed the existing lithography technology principles and manufacturing processes, only making slight modifications to the Chip design.

Even so, the performance of the Super Chip 1 had already surpassed the latest generation of Chips released by major giants in the United States by a level, and it even made the Huaneng Group see hope, directly contributing tens of billions in funding.

Now, Wang Shuo is prepared to use the technology that wasn't used in the Super Chip 1 for the Super Chip 2.

Taking a sip of water to rinse his mouth, Wang Shuo focused his gaze on the computer screen and began to think carefully.

Existing computer logic systems are all built based on hardware systems, so what is the most core part of a computer? Naturally, it is the Chip.

What is the most core part of a Chip? Naturally, it is the underlying layer.

The transistor is the underlying layer for all Chips in the world today.

Its function is very simple: it is a switch that cuts off or allows current to pass through.

But when someone defined the state where current passes through a transistor as 1 and the state where no current passes through as 0, and then combined this with Leibniz's calculation method, everything changed fundamentally.

The transistor leaped to become the founder of the computer world. Its operational logic directly determines the basic instructions and instruction sets of all Chips, thereby affecting computer assembly language and program execution logic.

This set of logic has ruled the entire information age until today, when the Super Chip 2 appeared.

Unlike the existing underlying layer of Chips, the underlying layer of the Super Chip 2 is entirely built from three-dimensional transistors.

This means that compared to general transistors which can only show two states—on and off—three-dimensional transistors can show one more state, reaching three.

With one more display state, the computer has one more option, moving from only being able to recognize 0 and 1 to being able to recognize -1, 0, and 1.

It sounds like there is no decisive progress, but if you were to place someone who truly understands computers in front of Wang Shuo right now, they would definitely exclaim: "Holy crap!"

Whoever it is.

Why was He Chenguang willing to throw away an annual salary of millions to run to Wang Shuo's newly established, penniless company to take an annual salary of less than 500,000 and work at The White House?

Because of three-dimensional transistors.

Why does Wang Shuo have the confidence that after producing the Super Chip 2, Ling will have sufficient computing power?

Also because of three-dimensional transistors.

But if three-dimensional transistors only simply improved efficiency, He Chenguang wouldn't have been in such a rush to resign.

The greater significance of three-dimensional transistors is their applicability to Artificial Intelligence.

To give a simple example, for the same program, in binary mode, the output result is only 0 or 1. Corresponding to logic, it can only be expressed as "true" or "false," with no third option.

This creates a problem: there is a discrepancy between human logical judgment in the real world and the computer's logical judgment, because for all problems, there are not just two answers—true or false—for humans; multiple situations exist simultaneously.

To match this discrepancy, computer engineers and software engineers need more time to design algorithms to achieve randomness in the user's perception, but even so, the essence remains unchanged.

But three-dimensional transistors are different. Beyond the two answers of true and false, they also simultaneously represent a third one, which is "don't know," represented by "0."

"-1" represents false, "0" represents don't know, and "1" represents true.

This is a logical thinking method that is closer to humans. This logical processing mechanism can allow engineers to create computer intelligence that was originally extremely complex or simply impossible to achieve using only simple code.

How could such a thing not make He Chenguang tempted?

Now, the basic technology is there, the manufacturing process is there, the equipment is in the preparation process, and the funding is basically sufficient. What remains is how to design a useful Super Chip 2.

This is also the reason why Wang Shuo spared no effort to attract He Chenguang.

As the development speed of Frontier Technology increases and the scope of operations grows larger, he cannot do everything himself. He needs people, especially those with outstanding learning abilities.

He Chenguang is like this, Wang Jie and the others are like this, Zhao Cheng is like this, and he isn't even planning to let go of those still-green juniors and seniors in school.

What? You won't come?

I'll give you a monthly salary of six thousand during your internship, and you won't come?

After becoming a full-time employee, it goes up to ten thousand directly, and you still won't come?

At the end of the year, I'll give you a bonus of at least thirty percent of your base salary, with no cap. Will you come or not?

The answer is almost certainly yes.

At least if this condition were given to Wang Shuo himself before he obtained the core, he would probably be secretly happy for a long time, and might even set up a table of food to invite a few friends to celebrate.

Shaking his head slightly, Wang Shuo said: "Ling, help me open the EDA design file submitted by He Chenguang."

Ling did not reply, but a new interface appeared in the window already opened on the computer screen.

Dense lines overlapped and covered each other, looking extremely complex.

This is the Chip blueprint designed based on three-dimensional transistors that He Chenguang conceived and provided to Wang Shuo. It uses ternary calculation, and the overall structure is much simplified.

Of course, He Chenguang also very humbly explained that this is just a preliminary plan, and the detailed plan is estimated to take at least half a year to be basically perfected.

To put it bluntly, this is just a prototype; it is still a long, long way from being ready to go directly into the production line.

But for Wang Shuo, it is already enough.

"Ling, screen the circuit, calculate the optimal path, and perform logical compilation."

"Okay, sir. Testing in progress, estimated time: 18 hours."

"Running this will occupy some resources and may lead to unpredictable results. Do you confirm execution?"

"Execute."

Wang Shuo picked up his teacup, took a sip, let out a breath of stale air, and said calmly.

His gaze inadvertently swept over the bottom right corner of the computer screen, where the time was clearly visible.

23:59.

Just as Wang Shuo's gaze swept over it, the time ticked over.

24:00.

A new day has arrived!