Interview with | Ding Jian, Head of GSR Ventures: AI investment pays more attention to scalability

"Such as technical barriers, moats are all to see, but to really do bigger and stronger, we must do scalability." Many companies want to do their own platform, but they don't want to really help the industry to achieve data standardization and promote the industry to improve efficiency. On December 6, Ding Jian elaborated on his "standards" for investing in artificial intelligence projects in an exclusive interview with reporters from Nanfang Daily and Nanfang + reporters.

Jian Ding, Managing Partner of GSR Venture Capital Fund

Jian Ding is currently the Managing Partner of GSR Venture Capital Fund, responsible for venture capital investment in the fields of artificial intelligence, big data, digital health, VR/AR and new media industries. Prior to joining GSR Ventures, Jian Ding was a co-founder of AsiaInfo.

Up to now, GSR Ventures has successfully invested in many enterprises such as Didi, Ele.me, horizon and so on.

Reducing human participation can amplify industrial efficiency

Southern Plus: How do you see the role of AI in "new infrastructure"?

Ding Jian: This problem should be looked at in reverse, AI itself is part of the new infrastructure, more importantly, AI is a purpose of the new infrastructure, one of the goals of the new infrastructure is to make analytical decisions with machine intelligence as the core.

How to understand this sentence, at the 2020 Yabuli China Entrepreneur Forum, Tian Xuning (chairman of Broadband Capital) joked, "Ding Jian said '5G is not for people'. I would like to add a word, 5G is not for 'individuals', but for enterprises, for the industry, for the industry." But I still insist on the idea that "5G is not for people", just as the highway is not for people to go. Do people need low latency and high bandwidth? Certainly not, 5G is mainly for machine communication and machine intelligence.

The most core purpose of new infrastructure is that people participate less and less in it, it is not difficult to imagine, in the industrial era what kind of factory is the most efficient, of course, fully automated, from beginning to end almost no human intervention, there will be such a miracle as "6 seconds off the line of a car"; then imagine, today's Taobao, Didi, if a process also needs people to tick, press the keyboard point to confirm, there may be such an efficiency as now.

Therefore, in my opinion, the new infrastructure needs to reduce "human participation" from it, and realize infrastructure construction and industrial upgrading with machine intelligence as the core. In this way, we can leverage the improvement of the efficiency of the industry by hundreds of times.

South+: Man is replaced, so where do people go?

Ding Jian: In fact, with a significant increase in productivity, eventually people's quality of life will be greatly improved. If it were not for the dramatic increase in efficiency caused by the Industrial Revolution, there might not have been a five-day workday. So it's the distribution system that needs to change, not the fear that AI will replace people. In the future, it is not necessary to work to have a source of income, machines can help create wealth, and people only need to work 2 to 3 days a week. People do more creative, valuable work.

Investment Opportunities Lie in "AI Middleware"

Southern+: How do you see investment opportunities in AI?

Ding Jian: Just like when we look at high-speed rail, we can't just look at how advanced the high-speed rail technology is and how fast the car can run, without thinking about what kind of rails and systems (infrastructure) are needed to make the train run at such a speed. By the same token, the current infrastructure is not enough to make artificial intelligence, big data, the cloud, etc. run.

For example, in the medical field, there is a small cycle, the first step is to collect data, the second step is the doctor's brain to analyze the data based on experience and knowledge, and the third step is to come up with a treatment plan and start over and over again.

In the future, this is changing, the first step of collecting data is no longer simply a CT, but may be a personal 24-hour health data collection, gene data collection, etc., the second step, so much data is no longer handed over to doctors to deal with, but machine intelligence to assist decision-making, bringing together a variety of data. The third step is that the doctor only needs to choose the treatment plan, which is a big cycle. In the future, every industry will move from a small cycle to a large cycle.

Among them, machine intelligence is not necessarily better than human decision-making, but it can make many decisions faster and at the same time, thereby greatly improving efficiency and improving scalability. So what we're focused on is the infrastructure that drives this big cycle, where the rails are. For example, some middle offices, middleware, operating systems, etc., they can allow enterprises to better apply big data, cloud, 5G and other technologies, and investment opportunities are also hidden here.

Southern+: When investing in the field of artificial intelligence, what characteristics or indicators will you pay more attention to?

Ding Jian: Individual areas can open up small cycles, but most of them cannot form a large cycle, so scalability is very important. The Internet has reached the global level of interconnection, but in the industrial field, enterprises in the data, applications, good can open up the upstream and downstream, poor even the internal information of enterprises is not interoperable.

South+: Many companies are also emphasizing building platforms and building ecosystems, is this different from the scalability you understand?

Ding Jian: The Internet giant in the consumer field itself is a super large platform. However, the industrial platform and industrial ecology are very poorly scalable. At present, in the industrial field, it is more like when the Internet has not yet been built, forming a bunch of information islands, data interaction, application interoperability, mutual recognition of results, etc., and its complexity is far beyond the Internet era. Therefore, if the next step of the industry is to truly improve, industry-level interconnection, middle office empowerment, and standardization of data exchange and sharing are very important core elements.

South +: Data difficulties are indeed more common in the industrial field. The medical industry, for example, involves sensitive and private data.

Ding Jian: This is a misunderstanding, you should want to help him do the data, rather than always running to people to ask for data, but with the data to do things, and then say that I want to subvert you, people are even more unlikely to give.

Whether it is an Internet company or an ICT enterprise, it is to help traditional industries use data, promote data standardization, and realize data interconnection between enterprises and enterprises, upstream and downstream of the industrial chain. Overall, the first is empowerment, the second is standardization, which is the most lacking in the industrial field, which is also what platform companies should help traditional industries to do.

[Reporter] Gao Xiaoping

【Planner】Chen Hanhui Cheng Peng

【Production】Southern Industry Think Tank

【Author】 Gao Xiaoping

Southern Industry Think Tank

Source: South+ - Create more value