After the layoffs, I became a tech godfather

Chapter 215: Taking the lead in future technology

EDA, electronic design automation tools, especially automation tools for the chip industry, like the chip industry, have long been controlled by a few Western companies.

Currently, except for the EDA used by Huawei, which was developed by itself after being restricted and whose specific level is currently unknown, more than 14% of other chip design tools above 90nm use EDA tools from Synopsys, Cadence Design Systems of America and Siemens of Germany.

China's BGI has risen strongly in recent years and has achieved great results in some fields.

However, in terms of digital circuit design, there is still a generational gap between us and these old Western EDA companies.

Of course, this is not just a software problem, because production tools must always be combined with the production process. The world's major chip manufacturers, TSMC, Samsung, and Intel, are not located in China or are not controlled by China.

If production is limited, then the design tools for production will naturally have no place to use and no room for development.

This is common sense.

As for Huawei, it remains a mystery how it managed to develop 14nm, 7nm and even higher-level design tools without the support of production lines.

Not to mention others, even Hao Cheng was curious about these questions. He also asked He Gang, and He Gang had only one sentence to answer: It’s spelled out!
It’s just one word: fight. Hao Cheng didn’t ask in detail, but he knew how difficult this process was just by thinking about it.

On the Internet, many people are criticizing Huawei's Striver Agreement, extreme overtime and other issues, but sometimes, many things really cannot be generalized.

Take Hao Cheng himself for example. His daily work now almost reaches the standard of 996. Sometimes, moving forward is not just a slogan. What he has to pay for it is endless hardship.

Some Western countries once made groundless accusations under the pretext of forced labor, but they never thought about their own history of making money like bandits.

Of course, some domestic companies are indeed unscrupulous, and overtime and overwork cannot be generalized.

The path that Baiju Technology is taking now is to enable most people to arrange their lives easily and comfortably.

Apart from exploring, all basic work is left to AI to complete. Even exploring can be done with the assistance of Xiaosha.

Just like this moment——

"Xiaosha has been committed to improving and upgrading the production process of the production line, and in this process, the data we obtain is detailed and accurate. Coupled with Xiaosha's learning level and industry fine-tuning capabilities, it can be said that we can always develop automated design tools that meet our own high-performance chip production lines.

"Xiaosha already has all the capabilities in this area. As for the human interaction interface, we have worked with Huawei and HuaDa JiuTian to develop two versions of commercial tools.

“It will also be completed in the near future. As for the specific progress and when we can use the EDA of these two companies, colleagues from the two companies will share the specific information with you at 9:00 and 11:00 tomorrow morning.

"I will stop here for traditional chip manufacturing and related arrangements." Hao Cheng said, "Let's take a break first. Twenty minutes later, I will continue to share with you the second half of the report on the core differences between metacognitive chips and our current traditional chips."

……

Hao Cheng took a sip of water and walked out of the lecture hall. He had just received a reminder from Xiaosha that Cao Song had contacted him. There must be some urgent situation that needed his instructions.

"Cao Song, what's going on?"

"Oh, don't worry about this kind of thing. We should collect the money we owe. At worst, we will pay for the penalty and then put the relevant companies on the blacklist. We and the companies related to us will never cooperate with them again. Even if we make a list later, we still can't punish them!"

"They didn't break the law, but we are a private enterprise. It's our freedom to choose our partners. It's not illegal to put them on the blacklist!"

"It won't affect our reputation at all. These bastards should have been dealt with long ago. Everyone will stand on our side. Do you believe it?"

"Okay, notify the personnel, administration and legal affairs department, and Xiaosha, and have them cooperate with you, that's it!"

Cao Song felt so relieved after hanging up the phone. He hated being morally blackmailed the most, especially by people who had no morals themselves. It happened that Mr. Hao didn't buy it, so he could do something big.

……

In the lecture hall, people were chatting in groups of three or five, exchanging information with each other.

"Mr. Zhang, your Moore Technology is going to make a fortune this time!" Several people around Zhang Jian were all envious.

Mr. Hao's sharing in the first half was all about graphics cards, 3D graphics cards, three-dimensional space computing cards, and immersive five-sense computing cards. The road ahead for Moore Technology is so bright.

Following the guiding light, Moore Technology will definitely become the best graphics card in the world.

What's more, now that we have a direction, design tools, and production problems have been solved, there will be no obstacles at all to move forward.

"Everyone is the same!" Zhang Jian was very excited, but he also knew that those present today were all strategic partners invited by Baiju Technology, and they all played a very important role.

It can be said that after coming here, everyone has automatically formed an industry alliance.

Zhang Jian said: "What's so good about traditional chips? They will only account for one-fifth of the market share in the future. Next, Baiju Technology will definitely help us to develop design tools, production tools and application chips for metacognitive chips, which are the big deal."

"So, what is the core difference between metacognition and current semiconductors?" This will be discussed in the second half, and at this time, some people began to think about this question.

"Not sure?" Zhang Jian also thought about this question, but he didn't figure it out. "But I think the word 'cognition' may be very important. We can verify it later."

Everybody was talking at once, their thirst for knowledge just like freshmen who have just entered college and, before the first class meeting, imagine their future employment and prospects in their major.

Yu Dong and He Gang obviously know more, because they have exchanged opinions with Hao Cheng many times on the cooperation of the three-dimensional data five-sense network, and the network and the chip, these two things actually complement and enhance each other to a certain extent.

After discussing it, Yu Dong also expressed his own feelings: "How should I put it? Let's put it this way. Ten years later, it will be a whole new world."

He and the Huawei team have been busy these past two days. Using limited information, they have produced a predictive scenario of what the world's technology will be like in ten years.

In response to the technology required for this scenario, they also released an internal white paper "Smart World 2035". They carefully evaluated the technical route of Baiju Technology, which parts of the route they could participate in, and conducted in-depth research.

Follow Baiju Technology and reserve relevant technologies in advance.

Mr. Hu Kun said at the last meeting: We do not seek to surpass Baiju Technology, but we must at least surpass all other competitors. In this way, once Baiju Technology needs assistance, it will definitely choose us.

This is absolutely true. Isn’t the cooperation on the three-dimensional data five-sense cognitive network based on Huawei’s achievements in the field of communications?

Although the current communication technology is far behind the five-sense cognitive notification, Huawei is already the best, so Baiju Technology will think of Huawei first. "Ten years later, I think your prediction should be a little more radical."

When several people were discussing this topic, Hao Cheng just finished a phone call and said, "But there are too many things to do. We really need to sort them out and prioritize them."

Hao Cheng read Huawei's white paper. The prediction seemed very radical to others, but to Hao Cheng, it was conservative.

Because it is predicted based on the topics disclosed by Baiju Technology, but in reality, it is just the surface.

"Even more radical?" Yu Dong was stunned.

"Okay, just keep predicting like this for now." Hao Cheng looked at the time and saw that twenty minutes were almost up, so he went on stage and prepared to start the second half of his sharing.

He Gang followed up with a sentence: "Boss Yu, I said at the meeting, predicting Baiju Technology? And predicting for ten years? Then the company must prepare this white paper and update it every two months."

"Maybe you really know more about Baiju Technology!" Yu Dong could only shake his head and sigh.

Others around me are a little numb - how can Huawei always be one step ahead in everything it does? What white paper?

"No wonder we have achieved such great success!" This is Zhang Jian's inner feeling: "We can always seize the initiative step by step. When we go back to Moore's Law, we must also make predictions and assessments in this regard and take a good position in the future technology route in advance."

……

Two minutes later, Hao Cheng returned to the stage, and the bustling crowd below instantly returned to silence.

"What is the difference between a metacognitive chip and a traditional chip?" Hao Cheng did not beat around the bush and went straight to the point: "It starts with a concept, which is the instruction set.

“The most basic operation and calculation unit of our traditional chips is the instruction.

“There are many types of instruction sets, including ARM, RISV, X86, and our LoongISA, which is an extension of MIPS, etc., and the instructions of computing power card devices are even more numerous. Each company has its own instruction system, just like NVIDIA.

“Traditional chips are all instruction-based. Remember the word ‘instruction’.

“As for the metacognitive chip, although I also call it a metacognitive instruction set, in fact, its essence has gone beyond the scope of instructions and has risen to the level of a [sensing-decision] mapping set.

"Some people may ask, since they are both collections, what is the difference between them? I would use one sentence to describe it: the difference is even greater than that between humans and dogs!"

"Pfft~" Many people laughed at once. They thought Hao Cheng was just making another metaphor, but unexpectedly, Hao Cheng really started talking about dogs, and the following explanation was refreshing to them. It was something like this -

Dogs cannot understand human words. The words here refer to language.

For example, if you say to a dog, "Let's go eat," it won't understand.

However, specific instructions can train dogs.

For example, every time you howl "eat", then feed it food. Over time, as long as you howl these two words, the dog will know it's time to eat.

Does the dog understand your language? No.

Even if you don't say the word "eat", but replace it with any other word, as long as you call it and feed it food, you can achieve the same effect over time.

On the other hand, if your command word is "eat", and you yell, "You are not allowed to eat today", the dog will still think you are going to feed him because the sentence contains the word "eat".

That's the instruction.

And what about metacognition?

That is equivalent to the revival of spiritual energy. The dog has mutated and can really understand what you are saying and what you mean.

This is when a dog truly learns a language, even if it is just one sentence.

At this time, the essence has already begun to change.

"Some people may have this question in their mind: Is the relationship between instruction and actually learning the language, or between instruction set and metacognition, a relationship of quantitative change leading to qualitative change?"

Hao Cheng took a sip of water to moisten his throat, then paused for a moment to give everyone present some time to think.

For example, the two words "eat" is a command. After training for a period of time, the dog will understand it.

"Squat down!" is also an instruction, and I gradually understood it!

"Scream!" I understood again.

When all the words have been made into instructions and the dog understands them all, does it then understand the language?

In fact, by that time, Hao Cheng didn't even need to answer, everyone understood what was going on.

Instructions need to be produced, but truly understood language does not require this.

The process of [sensing-decision-making] is a creative process, and also a process of simulating the formation of thinking. And every step in this process is related to cognition and thinking.

Instructions are completely different; they are purely mechanical and are fundamentally different.

At this time, everyone understood the logic of the metacognitive chip, and combined with what Hao Cheng had said before about the future pattern of the chip industry, the blueprint in their minds gradually became very clear.

At this moment, the large screen behind Hao Cheng returned to the horizontal line that was originally drawn.

He said: "This line is placed here. The part below this line is pure traditional semiconductors. This part in the future will be the basic plate of domestic chip design, production and application companies. Baiju Technology will not participate.

"For chips above this line, whether it is fusion chips or metacognitive chips, we will also choose strong partners to work with us to create this future blueprint.

"You can all think about what position you can occupy here and prepare in advance so as to gain an advantage in the future technological development."

(End of this chapter)