Wen 丨 Nanliang
Produced by丨Niudaocaijing
The battle of Tencent's big model was handed over to Tang Daosheng.
In the eyes of those familiar with Tang Daosheng, Tang Daosheng has been playing the role of Tencent's "pioneer", often taking over some things that others cannot do or are very difficult, and even going to wasteland that others are not optimistic about to explore, but this is an interesting intellectual game for Tang Daosheng.
Just like walking a maze, many people go from the starting point to the end point step by step, but if you start from the end point to the starting point, you will find that there will be much fewer forks in the road and the route will be clearer. Tang Daosheng and his children explained: "This is the way of thinking in mathematics, and the most particular thing is to start with the end. ”
But on the road to Tang Daosheng's To B business, there is no lack of stepping into a fork in the road.
Compared with Huawei, which established the "integrated" strategy early on, Alibaba Cloud is comparable to Tencent Cloud in "taking a fork in the road", but Alibaba Cloud has already achieved profitability one step ahead, how long will it take for Tencent Cloud?
Now that artificial intelligence has swept the industry, business is also facing innovation, which is both a challenge and an opportunity.
Not long ago, Tencent Cloud released a mixed element big model, but the move released after Ali and Baidu made the market disagree on the first mover advantage and the late starter, and Tang Daosheng did not think that it was necessary to "catch up", but the "maze game" of this large model has begun, and no one can say for sure how to go right.
1. The "maze" inevitably leads to forks in the road
Even if Tang Daosheng sees "starting with the end" as the least wrong way to solve the problem, it is impossible to avoid getting it right every time.
For enterprises that have been doing To C business, opening up the To B market is undoubtedly "crossing the river by feeling the stones".
"In order to seize the market, Tencent's early cloud business also had to take projects at a loss, did a lot of integration and turnkey business, took the lead in winning orders, and integrated self-developed products and third-party products." Speaking about the early operation of the cloud business, Tang Daosheng sighed: "In general, although these businesses have huge revenue, their costs are also high, and the profit margin is not high. ”
Cloud business is "difficult to make money" is the current industry generally default.
More and more players are pouring into the market, but the growth rate of the market is slowing down, and manufacturers have even started a "price war". Against this background, Tencent Cloud's domestic market share fell from 11.1% to 9.9% in the second half of 2022, and its ranking fell from second to fourth. In this context, the transformation to "be integrated" has become one of the choices of large manufacturers.
Ma Huateng once said frankly: "Tang Daosheng did not want to do integration before, but Tencent was often wrapped up in market share and public opinion, and was forced to do general contracting to gain market share." Under the situation that large manufacturers have turned to "being integrated", Tang Daosheng is also more determined to transform Tencent's To B business from "integrated" to "integrated".
Talking about the feelings of To B for five years, Tang Daosheng also said frankly: "This is a process of continuous exploration, trial and error, discovering problems and correcting. "It turns out that many people have business experience in To C, but there are still many differences in the To B track, and its business essence, logic, cost structure, teamwork model, etc. are different." The early development situation that emphasized scale over profit was also considered by Tang Daosheng to be the biggest trial and error.
It is inevitable to walk into a fork in the road in the process of walking through the maze, but in Tang Daosheng's opinion, this is not a big problem. "What is clearer is that we have been moving in the direction of solving problems, more and more to B." The big model that brings reform to the To B track is Tencent's hope of solving the problem.
The rise of large models will reshape Tencent's To B business, but how to leverage the real source of profit is the focus of Tencent Cloud's thinking, and it is also the test faced by Tang Daosheng. Of course, the industry changes brought by the big model not only test Tang Daosheng, but also test the entrants under many waves.
2. "Tencent Cloud" was born in doubt
Tang Daosheng is now known as the "father of Tencent Cloud", but he has been tested on the road to pioneering, and even if it is only to get Tencent to agree to start a cloud business, it is not easy.
Born in the golden age of Hong Kong, Tang Daosheng grew up to study in the United States, began to study computer engineering at the University of Michigan, and then went to Stanford University to study for a master's degree in electrical engineering. Before moving to Tencent, he worked at Sendmail Software, managing the R&D team to develop large-scale email systems and anti-spam filtering systems.
It wasn't until 2004 that Tencent rang the bell of going public and became a shining star in the stock market, and the following year, Tang Daosheng received an invitation from a Martin Lau, who is also a Stanford alumnus and is working for Tencent. At that time, Tencent urgently needed innovative technology bosses.
This year, Tang Daosheng officially joined Tencent, and at the same time, he also ushered in the bottleneck of the QQ space in Tencent's first battle. He thinks: "This is my battle of life and death", but at the same time, it is also his battle for fame.
At that time, QQ Space could not withstand more than 60W of simultaneous online people, and Tang Daosheng was not the first person to be sent to solve this problem, but he was the only one who succeeded. Under the continuous development of QQ, Guangdiantong and Tencent Cloud businesses have been incubated, but these two early days are not optimistic, and there are not many doubts about doing cloud within Tencent.
At that time, there were two opinions in the Tencent General Office. One believes that the cloud is unprofitable, costly, and a "hard job of digging coal", and even the president Martin Lau undecided. Another voice, including Tang Daosheng, believed that Yun could not be easily given up.
In order to keep the cloud from being abandoned, Lu Shan, president of Tencent's technology engineering business group, privately arranged for the team to support its research and development, and Tang Daosheng hid the cloud team in an open platform. But even so, in the face of unclear results, doubts from many parties and lack of resources, the pressure on the team and Tang Daosheng has been hanging high.
Until 2015, Ma Huateng proposed the strategy of Internet +, and even named Tencent Cloud as one of Tencent's key businesses in the financial report, recording an increase of more than 100%. This is the first time that Tencent has identified the importance of cloud business.
In 2018, Tang Daosheng was selected by Tencent to be in charge of the cloud business, which also forced Tang Daosheng to leave the To C field he was most familiar with and turn to develop the To B business. Although Tang Daoshen has always supported that cloud business cannot be abandoned, it is also a fact that there is no experience in the field of To B, and in the process of development, it is inevitable to enter the fork in the road of "integration".
3. The new "maze game" opens
Artificial intelligence is Tencent's hope to solve the problem of "taking a fork in the road", but after it released the mixed element model, there were many voices in the market who thought that Tencent was a little slow to enter the game.
Compared with the hybrid model that Tencent announced only in the second half of this year, Ali and Baidu released their own large models in the first half of the year. It is reported that in March and April this year, Baidu and Ali successively released Wen Xin Yiyan and Tongyi Qianqian, in contrast, among the three BATs, Tencent's large model was the latest to be released.
Although Tencent's hybrid model was released not long ago, Tang Daosheng believes that there is no "catch-up" for the backwardness. "The industry is too new, it's still in its early days, there's no market share or proportion, the big model is a marathon, and now it may only run a kilometer." Tang Daosheng further said, "Tencent only looks at itself and does not look at others when it makes a big model. The hybrid model has been tested internally within Tencent for a long time, and in terms of launch time, it has not considered racing with peers. ”
Ma Huateng gave an example: "For the industrial revolution, taking out the light bulb a month earlier is not so important in terms of a long time span." ”
Obviously, Tang Daosheng sees the heat of the industry in his eyes, but he is also clearer about his views. After all, although artificial intelligence has only caught fire this year, Tencent Cloud has long studied it.
In 2019, Ma Huateng said that Tencent had established four AI labs; In April 2022, Tencent disclosed for the first time the R&D progress of its hybrid AI big model; In December of the same year, the first low-cost, landable NLP trillion model in China was launched.
But in the current hot artificial intelligence, Tencent's opponents are also not weak.
The spectator erupts in with players, and this spectacular scene is even called "100-model battle". For the current situation of many players entering the big model, Tang Daosheng feels that it is the key to recognize the big model compared to earning sooner or later. "At present, many people know that the big model is a vent and want to seize this golden opportunity, but in fact, what problems to use the big model to solve, what value and how much value to obtain at what cost, in fact, have not thought clearly."
The goal of Tencent's big model is clear, that is, to "transform" thousands of industries. In June this year, Tencent Cloud announced its MaaS service solution for the first time: it has built an industry model store, covering 10 industries such as finance, cultural tourism, government affairs, media, and education.
In terms of direction, like Tang Daosheng's idea of starting with the end, the birth of Tencent Cloud big model is the first clear goal, which also makes Tencent's road to big model have a clearer direction.
Tencent believes that the big model is only the beginning, and the industry application is the future, which Ali and Baidu also have the same view, which is also destined to meet sooner or later.
The new "maze game" has begun, the major forces have entered the game one after another, Tang Daosheng, who is good at "starting with the end", how to break the game this time?