Looking at the evolution of technology from the "cold war" of chips

On November 10, the European Parliament passed the "EU Game Act" by an overwhelming majority of 560 votes. This means that the legislature has also officially recognized the huge value hidden behind electronic games.

Laurence Farreng, a member of the European Parliament and the main promoter of the resolution, said in the debate before the resolution: "For more than half of Europeans, video games are an important part of their cultural life". It is seeing the prosperous civil consumption market and potential behind the game ecosystem that the EU passed such a bill.

In fact, the opening and development of the civil market not only has a strong role in promoting the economy and culture of all countries, but also is particularly important in the scale of high-end and sophisticated industries represented by chips and artificial intelligence. I just recently read some books about the development of American chips and share them with you.

In 1955, in California, the famous "father of transistor", Nobel Prize winner in physics, MIT genius physicist Shockley and a group of young researchers were trudging in an "unknown place".

At that time, the initial generation of electronic computers led by the US Department of Defense had been completed and achieved remarkable results, but the computer just born was huge and did not meet the expectations of the Department of Defense.

The Ministry of Defense hopes to apply computers to a wider range of fields, such as the biological and aerospace fields. Almost all research institutes also hope to have one, and the military hopes to put simple control units (later chips) into missiles to make them more accurate. Even the radar and artillery are connected through chips. After the radar sees the target, the artillery can shoot directly without the gunner's adjustment.

But the computer at that time was too big to be used.

So several famous laboratories have carried out competitive research, including the British Radar Research Institute, Bell Laboratories, and the Shockley Laboratory mentioned at the beginning of the article, which he founded, are all doing research for this project. Shockley originally worked as a researcher in Bell Labs, and after he made achievements, he came out to work alone.

Compared with Bell Labs, Shockley Labs may know little about it. In fact, it has made a great contribution. However, the biggest contribution is that eight famous "traitors" came out later. Among them is the Moore in the "Moore's Law" we know well. He has been the technical director of Intel since then, and is still doing scientific research in this laboratory at this time. These eight employees do not agree with the founder's concept of Shockley.

As a scientist, Shockley has little interest in the market and demand. He only wants to do basic research, but has little interest in cost, mass production and profit.

But the eight people at hand think that the most important thing is to reduce costs, scale production, and develop integrated circuits, so that more people can afford these technologies. At that time, cheap control units can not only be stuffed into missiles, but also into radios and televisions, making computers smaller, more companies can afford them, and then the research institute can achieve profits.

This makes their leader and chief scientist Shockley feel very funny. As a scientist, what do you want to make money from basic research? So he not only rejected these people, but also said very rudely that since you eight want to work in the laboratory, you should listen to him well.

These eight people later left Shockley Laboratory and founded a new company called "Fairchild". The source of this name is that the father of the investor who invested them is "FairChild".

Some partners may not know this company, but in fact, this company is a milestone in the history of semiconductors and computers. It changed its focus on research and development to the mass production of integrated circuits based on silicon, and achieved great success. Within a few years, it swept its old owner into the historical garbage heap.

Later, the line of the management of Xiantong Company and their technical backbones collided again, and the backbones in charge of research and development fled again in large numbers. Like dandelion, a large number of semiconductor companies were established, including the later famous Intel and AMD. These companies all worked near Xiantong. Later, a large number of semiconductor companies were concentrated in that area, so they called it "Silicon Valley".

As soon as Silicon Valley appeared, it expanded crazily with the help of another new thing, that is, venture capital. After all, those entrepreneurs are basically scientists and college students. When they save enough money to start a business, the day lily is cold. And the trial and error cost has always been a big problem for entrepreneurs, and the emergence of "venture capital" has solved this problem.

And unlike previous investments, venture capital is not the end of the matter when investors give money. Instead, investors are deeply involved in the operation and use their contacts and relationships to help entrepreneurs win more resources.

Of course, the rapid progress is also related to their continuous flow of elite immigrants. For example, Intel founder Grove is Hungarian, Jobs' biological father is Syrian immigrant, Google Brin comes from a Jewish family in the Soviet Union, Nvidia's Huang Renxun and now AMD's Su Zifeng were both born in Taiwan, China, China, and Lin Benjian, who played a decisive role in the extreme ultraviolet lithography machine, is a Chinese Taiwanese born in Vietnam.

But this is not the most important thing. The most important thing is that at that time, the United States established a trading system, in which everyone had the money to buy these most advanced technology products.

At the beginning, the US military was buying these chips. Later, large companies and research institutes were also buying them in order to improve their work efficiency. Later, the software ecosystem was gradually enriched, which can be said to bring a large number of continuous technological needs and high returns to the chip industry, which also promoted chip research and development and broader applications in the reverse direction.

In fact, before the birth of Intel's 4004, the functions of the chip had been determined long ago. Now, only by changing the user instructions stored in 4002, can the 4004 achieve different functions, and an independent software industry was born.

At the same time as Intel launched the 4004, another computer giant, IBM, started to price software and hardware separately for the first time in the industry, which stimulated the development of the software industry in the commercial market dimension. It was at this time that large and small software companies were established. From 1970 to 1980 alone, the growth of the output value of the US software industry has outpaced that of the chip industry.

Here is another thing against common sense. The earliest large-scale application of chips is not computers, but all kinds of game consoles. It can be said that the application software represented by games has promoted revolutionary household products such as game consoles and chips beyond the understanding and imagination of most people, and even played a decisive role.

In fact, all the kids who have experienced this in the 1990s feel that the order of our contact with electronic products is also step by step learning machine, Super Mary Soul, then the Internet cafe, the Red Alert and the Frozen Throne, and finally the home computer.

The same is true in the West. When personal PCs first came out, they were also extremely expensive, but game consoles were much cheaper, so the game console industry became the main battlefield of the chip arms race. In 1977, the launch of the Atari 2600 caused a sensation, and the penetration rate of GB (Gameboy) in Japan was even higher. The sales volume was conservatively estimated to be more than 100 million units, and the following PS sold more. In order to attract users, these game manufacturers scramble to put the most advanced chips into their game consoles.

Even later, the popularity of computers was mainly related to games. College students knew that they bought computers for learning, and then they became game consoles. Now it is more obvious that most people buy desktops and smart phones. In addition to taking photos, video editing, and image repair, they also focus on game performance.

TSMC, the world's first chip generation factory, has 49% of its capacity to make mobile phone chips, and Apple is their major customer. The biggest driving force for chip progress in recent years is mainly the app for eating, drinking and playing on mobile phones. Among these apps, various mobile games are the most powerful and most powerful. A high-performance game phone is much more expensive than one that can't play games.

Moreover, it is particularly obvious in the game field that no matter how fast the hardware progresses, it will eventually be eaten up by more and more game software. The software will always be one step faster than the hardware, so we must frantically explore the hardware performance. In this arms race, AMD and Nvidia won. The founder of AMD is also from Xiantong. Huang Renxun of NVIDIA was an engineer of AMD.

Nvidia started its business by making video game cards. The revenue from games has always been their main source of income. In recent years, the proportion has been more than 40%, higher in the early stage, and 100% for a long time. Then NVIDIA took part of its profits and invested heavily in R&D, and gradually became the global GPU leader.

It is also because the parallel and complex computing power of the graphics card is getting higher and higher, which is just applied to AI. The highly configured version of Google's AlphaGo Distributed uses 1920 CPUs and 280 GPUs. Now the domestic training of AI models is mainly using GPUs.

In addition, who would have thought that today's UAVs, fighter planes and tanks are stuffed with graphics cards, especially military UAVs, which need to process a large number of images, are also stuffed with high-performance graphics cards.

In this process, in order to grab orders, those semiconductor companies spared no effort to reduce costs and improve performance. This unscrupulous competition made those chips faster and cheaper.

Low-priced products further make more people affordable, and more and more people join the buying force, further stimulating manufacturers to increase their research and development efforts.

The money spent by the masses to buy electronic products, after entering the technology company, the technology company will hire more excellent people, pay higher wages, and sharpen their heads to develop more advanced technology. College students can also go to these technology enterprises to do research and development after graduation, resulting in a huge middle class. The consumption capacity of the middle class can support further research and development.

This has formed the model of "consumption feeds back research and development, competition stimulates innovation", a business cycle of continuous expansion, once started, it can not stop.

This cycle is both a commercial feast and a cruel survival war. Numerous companies have been looking for a way to research and develop, and in the end, only a few solutions can be implemented, and the remaining companies will die.

Now you know that the chip-related companies, such as Intel, AMD, TSMC, ASML, Texas Instruments and Hynix, are all the dead of countless peers, and they still have to run crazily to prevent being surpassed again.

This pressure that will be eliminated at any time is the driving force of their research and development.

Previously, it was said that 5 nm is almost the technical limit. However, recently, Samsung has purchased a lithography machine with 2 nm energy and will put into production 1.4 nm in 2027. It seems conservative. The lemon of Moore's Law has not been completely squeezed.

Of course, the United States has not always maintained total victory. For example, in the field of storage, Japan, which has a more accurate grasp of the market, once grabbed the business of the United States. At that time, the storage in the United States was mainly used by large companies. Japanese entrepreneurs realized that home PCs were about to rise, so they concentrated on making hard disks for these small computers, and became bigger and bigger. After all, the market was close to limitless. After expansion, it began to devour the territory of American giants. Intel was almost forced to die, and finally gave up the storage transformation to focus on CPU.

It is also because of the large-scale use of cheap chips that the US military's high-tech weapons have reached a new level and fought a beautiful war of annihilation in the Gulf War. If those chips were not cheap like cabbage, they could not be used on a large scale.

In addition, many military technologies are not specially developed for the military, but are pulled from the civilian system, which can also reduce the cost of military research and development.

The most obvious is GPS. The US military has been brewing for a long time. Each soldier will be equipped with a GPS terminal. But the cost is too high, and satellites are secondary. Hundreds of thousands of GPS terminals are deadly. Until the price of civil GPS equipment has been reduced, this matter will naturally be solved.

Powerful civil technology companies can also customize products for the military. For example, many technologies of the US military are developed by IBM. In addition, because more advanced special chips have been developed and applied to precision CNC machine tools, these machines can produce more precise equipment, which can produce more advanced chips, forming an evolutionary closed-loop.

In contrast, the accuracy of Soviet weapons is not good. One of the most important reasons is that their chip computing power is not up, and the price is not down. Chips are not good, nor are related CNC lathes. They cannot produce anything more precise.

The reason is not complicated. They do not have a market elimination mechanism. They cannot quickly eliminate the dropped items, nor can they make the advanced ones stand out.

It was said that the Soviet Union did not develop integrated circuits in order to prevent them from being unsafe in nuclear war.

Not all right.

At that time, the specific direction of development was not planned, but developed. At that time, the United States did not choose one direction to move forward, but hundreds of directions were breaking through at the same time. However, in the end, silicon-based semiconductors gained decisive advantages, and other teams went to a dead end.

The computer predecessors of the United States and the Soviet Union all have the ability to let the younger generation listen to their own ideas. That's why Shockley cursed the eight men as traitors, and even threatened to make them completely unable to stay in the Jianghu. However, although Shockley had a great influence in the academic circle, few people paid attention to him in the market. Later, he became a joke, and the "eight traitors" became the god of Silicon Valley.

This was not feasible in the Soviet Union. Under the Soviet system at that time, there was no such thing as "working alone". Everyone had a certain hole. Unless he didn't want to get mixed up, he couldn't run away. The scientific research system of the Soviet Union has also seen such deviant rebels as the "eight traitors", but they can't go out alone, can only stay in the original unit, and finally completely disappeared.

In addition, the cost of modern scientific research is very high, and it is impossible to do anything without a large amount of financial support. Since the Edison era, invention and creation have been extremely costly. At the beginning, because of supporting Edison's career, the rich Morgan family was also full of contradictions. Little Morgan was very optimistic about Edison and invested a lot of resources, which made his relationship with his father very tense. Because the investment was too large, old Morgan could not understand Edison, and thought that little Morgan was just fooling around.

It can be said that the biggest problem in the scientific community of the Soviet Union is "authoritarianism", or "scholastic". After hard work, many pioneers in many fields have made achievements, and then become religious leaders.

This is similar in the world. It has been a common problem since Pythagoras executed his apprentice who dared to challenge himself. If they can, the elders will certainly suppress the younger generation, unless they really can't do it, they just can't do it under the market economy.

In addition, in the market economy, breaking the monopoly and barriers of others will gain excess profits, which will inspire countless people to engage in "disruptive innovation" one after another. Science and technology will move forward in the process of continuous self-denial. No matter how powerful enterprises are, they will be replaced by more powerful ones if they are not careful. No matter how advanced technologies are, they will be replaced by more advanced ones. No matter how powerful gods are, they will be replaced by new gods.

This is how progress comes about.

More importantly, the Soviet Union did not have a large consumer class to share the R&D costs. This led to the fact that the R&D expenses of the Soviet Union were financed, and the R&D expenses of the United States were ultimately shared equally by consumers.

The "original drive" of computers developed by the United States and the Soviet Union