China's tech giants need a deep rethink

Through the rise of ChatGPT, it is hoped that the Chinese technology giants will reflect deeply. In the future, there will be less short-term utilitarianism and exaggeration, more technological ideals and pragmatism in action, and truly return to the duty of a technology company, and steadily establish the core capabilities required for the long-term development of China's artificial intelligence industry.

ONE

Recently, the hottest topic of science and technology is undoubtedly ChatGPT, a chat robot product developed by OpenAI.

Before ChatGPT, in fact, domestic and foreign enterprises have launched some chat robot products, but none of them can be compared with ChatGPT. ChatGPT can not only answer questions and have long conversations with users on various topics, but also write poems, novels, scripts and computer code according to people's requirements. The quality of these content has reached an unprecedented new level, which can bring substantive help to people.

However, when the Chinese scientific and technological community marveled at the powerful functions and unlimited business prospects of ChatGPT, few people thought deeply about why it was born in the United States rather than first appeared in China.

After a systematic study of the context of ChatGPT, the author found that ChatGPT is essentially the product of the superposition of the technological idealism of a group of American entrepreneurs and the world-class AI industrial chain formed by a group of technology giants. Its birth in the United States is by no means accidental, and it is not easy for Chinese enterprises to replicate in a short time.

Among them, OpenAI is a non-profit organization with a strong sense of mission jointly established by many well-known entrepreneurs in the scientific and technological field, including Tesla founder Musk, president of Y Combinator, a well-known incubator, Peter Thiel, co-founder of PayPal, a global online payment platform, and Red Hoffman, co-founder of LinkedIn.

The original intention of its birth is that these entrepreneurs judge that AGI (General Artificial Intelligence) will inevitably appear in the future, and it is most likely to be born in DeepMind, a London-based neural network startup acquired by Google. In order to limit Google's dominance in the field of AI, they hope to create another AI organization that does not pursue profits as the goal to check and balance, to prevent the threat posed by unrestricted AI, and clearly put forward the mission of "democratizing the benefits of advanced AI", open all research and open all technologies.

The cohesion of technological idealism also enabled OpenAI to gather many top talents in the field of AI from the beginning. For example, Andrej Karpathy, the core executive who led Tesla to achieve outstanding achievements in the field of artificial intelligence, is a founding member of OpenAI. Recently, after becoming famous in the field of artificial intelligence, the super giant announced that he would leave Tesla and return to OpenAI.

In addition to entrepreneurs in the science and technology industry, OpenAI has also received strong support from Microsoft, Nvidia and other technology giants in the fields of capital and computing power during its development.

Readers familiar with AI industry will understand that there are three key elements in AI industry, one is algorithm, one is data, and the other is super computing power. In particular, generative AI products like ChatGPT rely more on AI computing power.

As the world's top cloud computing giant, Microsoft has the world's leading data center layout, and has established a large and costly AI dedicated chip cluster to promote its progress in the field of language model development. These data centers and dedicated chip clusters provide powerful computational support for the in-depth learning and training of multiple AI models under OpenAI. In addition, in July 2019, Microsoft also made an initial investment of $1 billion in OpenAI and became the "preferred partner" for the commercialization of OpenAI technology.

Behind Microsoft's super AI computing power is the world's leading high-performance AI chip of Nvidia, a GPU giant. At present, almost all of the in-depth learning programs in the United States are completely dependent on the accelerated computing of Nvidia GPU. Microsoft and Nvidia have cooperated for many years to jointly build the "world's most powerful AI supercomputer", which aims to deal with the huge amount of computing required for training and expanding AI.

Not long ago, Nvidia also designed the world's first single-box deep learning supercomputer, NVDIA DGX-1, specifically for OpenAI to deal with the most difficult AI problems. High-performance products such as NVDIA DGX-1 are missing in China and have also been banned by the US government.

With a major breakthrough in technology and products, Microsoft announced in January 2023 that it would invest $10 billion in OpenAI with a valuation of $29 billion. At the same time, its browser, search engine Bing, Office and cloud computing platform Azure and other product lines will deeply integrate ChatGPT.

In this way, on the one hand, all products under Microsoft will benefit from the strong AI capability support of OpenAI and achieve the improvement of product competitiveness; On the other hand, OpenAI will also benefit from Microsoft's huge customer base and rich application scenarios, form an increasingly strong snowball effect on technology and products, and build a more advanced intergenerational advantage over competitors.

It is also worth mentioning that OpenAI is still committed to maintaining its original non-profit intention despite accepting a huge investment from Microsoft. The investment agreement between OpenAI and Microsoft shows that the cooperation between Microsoft and OpenAI is divided into three stages:

In the first stage, Microsoft has invested 13 billion dollars in total and obtained about 49% of the shares of OpenAI. Microsoft has the right to obtain 75% of the company's future profit share until it recouped the investment of 13 billion dollars, after which the share proportion gradually decreased.

In the second stage, after Microsoft's profit reached 92 billion US dollars, the share ratio will be reduced to 49%, which is the same as the share ratio.

In the third stage, after Microsoft's profit reaches 150 billion dollars, it and other investors will donate their shares to the OpenAI Foundation for free, so that OpenAI can return to non-profit organizations.

TWO

Before ChatGPT became popular, in fact, a large number of Chinese technology enterprises had claimed that their AI level had reached the world's leading level. However, due to the fact that enterprises do not see much difference in basic applications such as visual recognition, voice recognition and content recommendation, and no enterprises show the particularity of "people without me", it was difficult for the general public to really identify the true and the false, and the advantages and disadvantages.

After ChatGPT became popular, some Chinese science and technology enterprises also announced that they had the ability to develop products similar to ChatGPT, and began to develop products and prepare for launching. For example, Baidu officially announced the ChatGPT-like project "ERNIE Bot", which is expected to complete the internal test in March; It is said that Alibaba Dharma Academy is also developing ChatGPT-like dialogue robots, which have been open to internal employees for testing.

The positive statements of Chinese science and technology enterprises can not dispel the public's doubts about their technological capabilities, but rather they will be considered as speculation. However, the true level of these enterprises will soon be revealed. Because ChatGPT provides a reference for these AI enterprises. Are they Internet giants like Baidu, Tencent, Alibaba and Byte? Is it a professional AI enterprise like Shangtang, Yuncong, Kuangshi, Itu and iFLYTEK? Or can smart hardware enterprises such as Xiaomi, Huawei, OPPO and Vivo develop products comparable to ChatGPT? The final product experience and user's word of mouth will not lie, but will explain everything.

However, the author is not optimistic about the commitment of the above enterprises.

In the United States, so many technology giants such as Google, Apple and Microsoft have launched similar chat robot products, but the only product that is really widely recognized by users is ChatGPT. This is because only OpenAI technology has broken the critical point of commercial use. At present, it is not difficult for domestic enterprises to make products similar to ChatGPT. The difficult thing is whether the content results generated by the products can have real value and be recognized by users, rather than generating some content that is full of flaws and makes people laugh.

From the perspective of the development law of the technology field, it is impossible for domestic technology enterprises to produce products comparable to ChatGPT at present. If OpenAI's technology is copied at such a low threshold, there will be no phenomenal popularity of ChatGPT today.

THREE

It is difficult for Chinese technology enterprises to take the lead in making products like ChatGPT that have made great breakthroughs in technology. Another fundamental reason is that the Chinese technology giants represented by BAT are gradually becoming mediocre. They are even difficult to be defined as technology enterprises. The business focus is only the traffic giants that obtain traffic through some technical means, and then realize the traffic through simple models such as advertising and financial lending.

When these enterprises have been living in such comfortable areas, revolutionary innovation is also out of the question. The appearance of ChatGPT has revealed the true face of Chinese science and technology enterprises that have always claimed that their AI level has reached world-class, and made the public truly aware of some huge gaps between them and their overseas counterparts.

To some extent, the current AI industry in China and the AI industry in the United States are essentially two different realms. China's AI enterprises only serve some very simple basic application level of explicit demand, with low threshold and homogeneity, while breakthroughs in more advanced fields not only lack the support of a large number of key elements at the technical and cultural level, but also lack the imagination for the long-term blueprint of AI industry. In the United States, the biggest achievement is not OpenAI and Google Deepmind, but the vision of technology leaders for the future of AI industry, and the strong AI industry chain built by a group of technology enterprises such as Microsoft, Nvidia, IBM, Tesla and Paypal under this blueprint.

Through the popularity of ChatGPT this time, I hope it can cause profound reflection of Chinese technology giants. In the future, there will be less short-term utilitarianism and grandiosity, and more scientific and technological ideals and practical actions, so as to truly return to the responsibility of science and technology companies and firmly establish the core capabilities required for the long-term development of China's artificial intelligence industry.

#semiconductormanufacturing?#semiconductorindustry?#offer?#oems?#costdown?#shortages?#distributor?#excessstock?#excessinventory?#consignment?#analogdevices?#texasinstruments?#automotiveelectronics?#industrialautomation?#iotsolutions?#pcbassembly?#pcbmanufacturing?#opportunity?#sourcingandprocurment?#engineering?#electroniccomponents?#automotiveengineering?#automotivetechnology?#automotiveindustry?#automotive?#automation?#global?#sourcing?#hongkong?#supercomputer?#digital?#edgecomputing?#electronicsmanufacturing?#semiconductors?#procurement?#cybersecurity?#machineslearning?#hybridcloud?#ai?#server?#computing?#artificialintelligence?#healthcare?#nvidia?#tesla?#processors?#cloudserver?#cloudstorage?#datacenters?#workstations?#network?#amd?#growthmindset?#storage?#manufacturing?#cloudservices?#networking?#intel?#memory?#semiconductor?#supercomputers?#cloud?#technology?#innvovation?#gpu?#electronics?#cpu?#servers?#supplychain?#iot?#leadership?#deeplearning?#bigdata?#cloudcomputing?#ems?#smartphone?#itad?#5g?#datacanter?#Components?#Semiconductors?#ICs?#CPU?#MEMORY?#SSD?#GPU?#HDD?#MCU?#EOL?#Inventory?#Excess?#Slowmoving?#Obsolete?#Management?#semiconductor?#testing?#qualityassurance