What is hindering domestic teams from researching products like ChatGPT?

Firstly, it is the scale of business. Due to geographical and cultural reasons, most domestic companies encounter some difficulties in going global, mainly in the domestic market, which is much smaller than the European and American markets. The same is true for public clouds, where the revenue and market value of AWS, Azure, and Google Cloud in the European and American markets are higher than those of Alibaba, Tencent, and Huawei Cloud in China. Since the development cost can basically be shared, the average salary of developers in American companies is higher than that in China, which can hire relatively more excellent talents; it can also generate more profits to support relatively long-term research, such as OpenAI, Deepmind, and Microsoft Research. Breakthrough innovations like ChatGPT are hard to come from product departments with intense development rhythms, they usually come from research departments without much short-term commercial monetization pressure.

Secondly, it is architectural superstition. Many domestic companies have been in a state of catching up for a long time and are accustomed to the copy/follow thinking mode. A paper published by Google or Microsoft is regarded as the Bible, but the same theory proposed by experts within the company is ignored. If ChatGPT does not support Chinese, I believe that a bunch of “Chinese versions of ChatGPT” will emerge in a few days. Architectural superstition is manifested as a preference for incremental innovation by patching up the existing architecture. When facing new architectures, new concepts, and new theories, there are often “three soul-searching questions”:

• Which customer can use what you are doing?
• How much money can it make for the company?
• Why didn’t others think of it?

Of course, ChatGPT is still a technology that is relatively easy to productize and can answer these “three soul-searching questions”, but other projects may not, such as playing Go, can AlphaGo make money? Projects like AlphaGo may not even be approved.

Thirdly, it is ecological superstition, which is manifested as believing that other people’s things are the standard, must be compatible, dare not propose their own standards, and dare not establish their own ecology. Sometimes it goes to another extreme and simply creates a closed system, promoting it through administrative orders or spending money. It is meaningless to blindly challenge the ecology that others have already established. We need to dare to establish a new open ecology in new scenarios and new architectures.

Architectural superstition and ecological superstition are partly because we are still relatively backward as a whole, and another important reason is the lack of adventurous spirit in our culture. As you can see, whether it’s Jensen Huang, the promoter of AI hardware, or Elon Musk, who always thinks about colonizing the universe, there is an adventurous spirit in their bones. This is also the spiritual power behind the European Age of Exploration.

Finally, I believe that none of the above problems will prevent us from conquering the stars and the sea. Six years ago, when I published my first SIGCOMM (a top conference in the network field) under the guidance of my mentor, almost the only ones in China who could publish SIGCOMM were Microsoft Research Asia. I would never have thought that now there are several SIGCOMM papers in China every year, and companies like Alibaba, ByteDance, Huawei, and various universities and research institutes have achieved a surge in top conferences. This is backed by the huge scale of network services and cutting-edge network research of major companies, as well as the world-class research talents introduced by major universities. In 2019, SIGCOMM was held in China for the first time, which can be considered a “turning point”.

ChatGPT, in my opinion, is the first AI dialogue system that doesn’t seem like artificial stupidity. It can also be said to be a “turning point” in AI dialogue systems. We are not short of big data and big models. As long as we break the architectural and ecological superstitions and encourage the adventurous spirit, I believe that China’s OpenAI and Deepmind are already on the way. I look forward to the arrival of this turning point.

Thinking is the problem; doing is the answer.