Chapter 1109 AI Bionics

Regarding the mobile phone business, Zhou Buqi's initial idea was to use the mobile phone as a terminal extension of Ziweixing's Internet business, which is a deeply tied relationship.

Through the mobile phone terminal, to forcefully promote Ziweixing's Internet business.

For now, no need.

In the era of PC Internet, Ziwei Star has achieved one of the best levels in China, and the era of mobile Internet will be bright.

In this way, there is no need to forcibly tie the mobile phone to Ziweixing's chariot.

You can spread your wings and fly freely.

The first is to change the name.

Zhou Buqi's suggestion was to change the name to "Esther". After discussion, Zhou Shaoning and Yan Guangming decided to call it "Esther", starting a new chapter.

Especially after Li Yinan joined the job.

Shen Xiangyang is very curious about this person. He has been sending people to follow up the whole process in the past few days, and then report to Boss Zhou.

"This person... how should I put it, he has an idea."

"Huh? What do you mean?"

Shen Xiangyang said: "He has a lot of ideas, and he doesn't like to express too much."

Zhou Buqi smiled and said: "Yes, it's normal to have a weird personality. What is he doing these days?"

Shen Xiangyang thought for a while and concluded: "Probably three things. First, he took the internal Aster mobile phone and compared it with the iPhone. Second, he went to Tsinghua University's Institute of Microelectronics to investigate for half a day with Xu Liangjie. ; Third, spent a whole day in our artificial intelligence laboratory."

The first two are easy to understand, but the last one is very strange.

For now, the mobile phone business has nothing to do with artificial intelligence. It will take at least three or four years before the Rokid smart assistant logs into the mobile phone.

Zhou Buqi asked strangely: "Is he interested in artificial intelligence?"

"I don't know."

"ah?"

"He is a weirdo, like a gourd. The day before yesterday in the artificial intelligence laboratory, he pulled many experts to ask about topics and directions in the field of artificial intelligence. I asked him what he was thinking, and he didn't say anything."

Zhou Buqi smiled, "It's still the same sentence, don't be afraid of doing wrong things, just be afraid of not doing things. No matter what he is doing, as long as he has ideas and is doing things, then he can do it. What I am most worried about now is that the salary is too high , so that he has no motivation to move forward."

Shen Xiangyang shook his head, and said confidently: "That's not true, he's a techie. His wife manages all the money at home, just a T-shirt and big shorts. Put it outside, no one will know it's a A billionaire."

"Well, that's fine."

"However, I can vaguely guess what he is going to do."

"oh?"

Zhou Buqi's eyes lit up.

This is not surprising, Shen Xiangyang is also the top scientist in the world, especially in the field of artificial intelligence, not to mention domestic, even in Silicon Valley.

Shen Xiangyang said in a deep voice: "The domestic chip technology is too poor. The chips we design independently are at least three or four generations behind foreign countries. They are all semi-public designs eliminated by foreign countries and passed the blockade restrictions. You let He made a general-purpose chip, and he directly took a big step to overtake the car in a corner to make a 4G chip, which is too difficult and the probability of failure is extremely high."

Zhou Buqi sighed, "There is no other way. No matter what kind of company it is, survival is the most important thing. 95% of domestic technology companies are struggling with life and death. It is difficult for them to even survive. Let them invest It is unrealistic to spend money on such high-risk R\u0026D.”

In the past two decades, technology companies have been regarded as the industry with the highest elimination rate in China.

In the beginning, they all had a state-owned background, and there were many obstacles. Except for a very few people who came out, most of them collapsed. For example, the Panda mobile phone and Kejian mobile phone sold by Xiaonei.com before, these are all products of state-owned enterprises, but they all failed later.

At present, domestic technology companies in the civilian field are basically dominated by private companies, and a few companies that have completed in-depth share reforms from state-owned companies are almost the same as private companies, such as Lenovo, ZTE, and Xunfei, including Haier, Gree, and TCL. An outstanding major appliance company that has come out.

It's genetically determined.

The attributes behind the positions in the system are low wages, high benefits, stability, ease, and iron rice bowls. The characteristics of these positions are completely contrary to the field of science and technology.

Being in the technology industry is bound to be accompanied by high salaries, high risks, and layoffs at any time. Even if they are all assembled, Bird's mobile phones are more exquisite than Kejian's mobile phones.

Aspiring scientific and technological talents, no one likes to retire in the system.

There is an even more important point.

If you go to a private enterprise or a state-owned enterprise with deep share reform, you can get more shares; in a traditional state-owned enterprise, you have nothing to do with shares. These days, it is impossible to achieve wealth freedom by wages, only stocks can.

This will inevitably lead to a phenomenon that technology companies are short of money.

A few college students want to do some technology research and development, how to do it? It is impossible to find venture capital, and R\u0026D is a bottomless pit. You can only apply with the state's industrial support funds.

However, young people in the technology industry are basically ignorant of the world and not tactful. Without background connections, it is almost impossible to apply for funds.

Compared with the rapid development of all walks of life, the profit rate of the technology industry is the worst. More than 95% of the technology industries are hovering between life and death. Only those 5% of the leading industries are shining, reaping great success, great wealth and great social influence.

If these 5% of the companies do not do high-tech research and development, it will be really useless.

Zhou Buqi was very firm, and said loudly: "Whether it succeeds or fails, Ziweixing must do it! R\u0026D is like a sea of ​​stars, let me rush!"

Shen Xiangyang shook his head, "R\u0026D is not a sea of ​​stars. If it were a sea of ​​stars, wouldn't everyone rush into it? Is there any need to call? R\u0026D is an abyss of hell, a dark world full of unknowns. "

Zhou Buqi said: "Even if it's the abyss of hell, let me go in! If there are more people walking, there will be a way. If the strong ones don't rush ahead, they will never be able to pass through hell and open up the road. The road between earth and heaven!"

Shen Xiangyang smiled, "Don't get excited, I know that your will to invest in research and development is very firm. I just want you to have a correct understanding so that your strategic decisions can be more accurate."

Zhou Buqi raised his eyebrows, feeling that there was something in his words, "What do you mean?"

Shen Xiangyang said: "This is also based on the analysis of Li Yinan's itinerary in the past few days. It is too risky to overtake on curves with 4g chips. It is impossible to overtake on straight roads, let alone curves. Eighty-nine times the car will overturn. There is a better idea, which is to change lanes and overtake.”

"Change lanes? How?"

"Do you know bionic chips?"

"Apple?"

Zhou Buqi had a foresight. He knew that Apple's self-developed A-series chips were called bionic chips.

"Apple?" Shen Xiangyang was stunned for a moment, "Is Apple going to make a bionic chip? Haven't you heard of it?"

Jobs believes that if you want to truly achieve the ultimate in software, you must "integrate software and hardware", and the core is of course the chip. With the success of the original iPhone, Jobs' confidence increased. This year, Apple made a big move. It spent 200 million US dollars to acquire the well-known chip company P.A Semiconductor, and immediately had 150 top chip engineers in Silicon Valley, starting the chip road.

This is the advantage of the Silicon Valley giants.

They can easily complete the integration of resources, talents and technology in Silicon Valley, and quickly own top technology across the track. The domestic technology giants are not good enough, and they need to explore a little bit. It is impossible for the authorities to approve it if you go to Silicon Valley to spend money to buy it, so you can only go to Europe to find a way.

Zhou Buqi said helplessly: "No, what exactly is the bionic chip you mentioned?"

Shen Xiangyang said: "A bionic chip is a chip that has added functions and behaviors that mimic biological systems. It is equivalent to having a certain amount of artificial intelligence, and it can also be called an AI chip."

"Oh, that's how it is!"

"Actually, it's mainly a neural network engine. Well... Let me put it simply, it was a scientist in the United States in 2006 who was also a good friend of mine. He proposed a restricted Boltzmann machine model and a deep belief network, which successfully Training a multi-layer neural network solves the local optimal solution problem of the backpropagation algorithm, which is the deep learning that our artificial intelligence laboratory is doing. With deep learning, there is the possibility of large-scale deep neural network learning. The first company to follow up was Nvidia, which developed the world's first AI chip..."

"Wait!" Zhou Buqi interrupted him, "What exactly is an AI chip? The 4G chip we are going to make is a general-purpose chip, right? What's the difference?"

Shen Xiangyang said: "The processing power is stronger, and the computing power is stronger! A general-purpose chip means that it can be used universally. The scope of application of AI chips is very small, mainly artificial intelligence. For example, yellow map identification, this is artificial intelligence. It seems simple, but It needs super powerful computing power. General-purpose chips can’t do it, and AI chips are needed to handle it.”

"Well……"

"Let's put it this way, do you know why Photoshop doesn't have a mobile version?"

"Because the computing power on the mobile phone is not enough?"

"That's almost what it means." Shen Xiangyang nodded, "When you open a large-scale game on the computer, if the processor performance is not good, it will be very laggy. The same is true on the mobile phone. For general games and videos, a general-purpose chip is enough But when encountering a huge amount of calculation, it is very difficult, such as video rendering, which cannot be completed by general-purpose chips. Well, in fact, we also have a solution, which is the cloud computing platform. If mobile phone users want to render videos in depth, they can upload the video Go to the cloud, and then let the processor in the cloud do the calculation and processing. But this is slow, and the process needs to be uploaded and downloaded, which is very troublesome. If you have a powerful AI bionic chip that can handle super-large calculations, you can directly In-depth video rendering is carried out. Another example is the face recognition on the mobile phone you mentioned before, which requires powerful artificial intelligence computing power, and general-purpose chips will be very difficult. At most, some simple face recognition can be done. AI chips can recognize more comprehensive and accurate."

Zhou Buqi nodded slightly, basically understanding.

He doesn't understand what the hell an AI chip is.

However, he has been in Ziweixing for so long, and he has a certain degree of technical cognition, combined with foresight, he can have a keen judgment.

The general direction of future technology is centralized cloud computing. Then, there is decentralized edge computing.

Probably the logic is that the technology is underdeveloped now, and the processing power of the local chip is not enough. What should I do if I encounter a problem that cannot be solved? Upload to the cloud to use cloud computing. Then feed back the results through the cloud.

When technology develops in the future, the local chip processing capability will be enough, and some complex processing does not need to be uploaded to the cloud, and calculations can be performed locally or at the edge near the local end.

The concept of this AI chip is in line with the trend of decentralization!

There is one benefit of decentralization - security.

Sending information to the cloud for processing is equivalent to handing over one's own content to the cloud computing platform, which might lead to leaks. It is safer to process directly on the local side.

Coupled with the fact that the bionic chip of the iPhone will kill all directions in the future, it can be seen that the bionic chip will come in handy.

"Well, this direction can be!"

Zhou Buqi soon had his own judgment.

Shen Xiangyang felt that this matter was very important, and further explained, "If smart phones develop further, they will inevitably handle more and more tasks on the mobile side. Calculations involving artificial intelligence are often very complicated, which requires AI chip."

Zhou Buqi narrowed his eyes, "So you think that Li Yinan's weirdness these days is also because of this track?"

Shen Xiangyang nodded and said: "It is very possible that he went to the Institute of Microelectronics of Tsinghua University, and his main research is the semiconductor major. He is also so interested in our artificial intelligence laboratory. It is very likely that he has ideas in this regard."

"What do you think?"

"hehe."

Shen Xiangyang smiled.

Zhou Buqi laughed and said, "What are you laughing at?"

Shen Xiangyang was silent for a while, and said slowly: "It's hard to say. AI chips are a brand new field. The foundation of the neural network was only built in 2006, and there are still too many technical options. It seems that there are hundreds of There are ways to go, and no one can say for sure which one to choose. The direction of AI chips is very good, but it is difficult to say what kind of technology choice it is.”

This is very helpless.

Zhou Buqi is very clear about the business direction, and AI chips will definitely become the mainstream in the future. However, it is not clear how to choose the technical direction of designing AI chips. Once you go the wrong way, all your previous efforts may be wasted.

However, this is also an advantage.

That is, in the field of AI chips, domestic and foreign countries started at the same time. As long as Ziwei Star enters the market in time, even if the development speed is not as fast as that of Silicon Valley, at least it will not be as big as the gap of three or four generations behind in the field of general-purpose chips.

Zhou Buqi said in a deep voice, "Old Shen, I think this idea works!"

"oh?"

"As you said, overtaking on a curve is easy to overturn, so it is better to change lanes and overtake. It is necessary to enter the field in advance to deploy AI chips. Well, at least you can walk on two legs. On the electronics side, let Li Yinan make a 4G general-purpose chip. We Organize a team here to try to do AI chip research, ten years as one day, you can take your time."

"Really?"

Shen Xiangyang was a little surprised, feeling that this boss Zhou was really impulsive and decisive.

Zhou Buqi sighed, "How can you not do it? Look at those domestic counterparts, everyone doesn't do it, how can our country's chip engineering make progress? Someone has to stand up. Old Shen, you go to Li Yinan , talk to him well, and then send me a detailed report."