First of all, I would like to give you a good old life.
At the end of the year, major technology manufacturers and research institutions have released their own research and judgment of the top ten technology trends in the new year. Of course, these contents are very beneficial, and I believe that readers and friends who love technology should also pay attention to them.
But to put it bluntly, if you look at the top ten trends, you can't help but find two problems, one is too narrow and the other is too fast. The so-called too narrow is the top ten trends released by manufacturers, which are often aligned with their own business scope. Of course, there is nothing wrong with this, after all, companies can't go to the trouble of advertising to others. But it is inevitable that there will be a suspicion that "what I do is the trend, and what I don't do is nothing". Too narrow a perspective can easily ignore the openness and diversity of the technology industry.
The so-called too fast mostly comes from some scientific research institutions or cutting-edge technology research and development departments of enterprises. To show their foresight, the trends they publish tend to be too cutting-edge, even a bit sci-fi. The frontier makes people wonder if they have traveled from 2024 to 2042.
This little question has been lingering in my heart for a long time, and it can't help but make people eager to try. As a result, there is this "Year of the Dragon, Top Ten Trends in China's Science and Technology" released by the brain-polar body non-blockbuster. There are three issues that we hope to address in this. One is to be compatible with multiple industry trends such as artificial intelligence, ICT, and smart terminals, and try to broaden the observation horizon of the technology industry. The second is to pay attention to the recent changes and opportunities in front of us, after all, the sea of stars starts between the grass and trees. The third is to be real enough, so that every trend has a basis to follow and a place to fall.
In our view, the background color of China's science and technology is open and solid.
The change of science and technology is the beginning of the tiger. This Jiachen year, we are continuing to play music and dance.
Trend 1: AI-native applications are emerging from the cocoon
Throughout 2023, it seems that the industry is discussing whether to make AI large models or AI native applications. Actually, this discussion doesn't make much sense. When large models become popular, everyone naturally wants to make large models, and once there is a popular AI native application, the focus of the industry will naturally shift to the application layer.
This is a question of business risk aversion, and it is also a question of heat.
At the beginning of 2024, we will find more and more unique AI-native applications around the world. From AIGC images, to personal assistants, AI games, and more. AI applications are like a pot of hot water ready to boil, and the next step is to boil after seeing the bubbles float.
This means that in the coming year, there will almost certainly be a boom in AI native applications, which will set off a new business outlet, and there should even be more than one AI application that will be taken seriously.
The replication and localization of these applications will be the focus of China's AI industry in the coming period.
Trend 2: Build a GPTS-like large-scale model development base
When it comes to AI technology, everyone should know about deep learning development frameworks. The principle of the framework is to platform the convergent parts of the algorithm into a general development base. After the explosion of AI large models, the development mechanism of AI applications is further evolving. In fact, GPTS can be seen as such an attempt. It uses the large model itself as a development mechanism, so as to achieve simpler or even no first-class AI application development.
Seeing that some people are discussing, can China make GPTS? This question is as pointless as there was a discussion about whether China can make a big model at this time last year. We have repeatedly emphasized that as far as the AI industry is concerned, there is only a capability gap between Chinese technology companies and the world's top level, and there is no generational gap. What European and American AI companies can do, China's AI industry will definitely be able to do.
GPT-like development mechanisms and more AI development platforms based on large models will be the focus of the industry in the future.
At the same time, from AI developers to large-scale model developers, this new identity and new industry will rise. Gathering the ecosystem of large-scale model developers will also become the layout direction of AI enterprises.
Trend 3: Pure-blooded Hongmeng, big points
On January 18, 2024, Huawei announced the launch of the "pure-blooded HarmonyOS" HarmonyOS NEXT, which is no longer compatible with Android. No matter how hard-mouthed the singers are, the matter of "Hongmeng has become" is firmly becoming the consensus of more and more people.
At this stage today, pure-blooded Hongmeng has a series of benefits. For example, its technical characteristics are deeply rooted in the hearts of the people, and there are still no competing products; For example, with Huawei's return to the terminal market, HarmonyOS applications have ushered in huge business opportunities; Another example is the emotional value and emotional value attached to Hongmeng, which is difficult to replace.
The benefits of technology, business, and ** will promote the development of pure-blooded Hongmeng. On the other hand, the argument of the declining singer is still "no one has done it before, so Hongmeng can't do it", which really seems a bit thin.
Therefore, we are sure that the next year will be a great year for pure-blooded Hongmeng. HarmonyOS's own technological iteration and third-party application ecological construction will be two changes worth paying attention to in the smart terminal market.
In the B-end market, the performance of open source Hongmeng is also worth paying attention to. Large-scale industry HarmonyOS applications have begun to appear, and industry HarmonyOS standards will be introduced one after another. We have a prediction: an important tributary of China's industrial intelligence will be industrial Hongmeng.
It's just to do something that no one has done before, that is, to see the clouds during the day and raise the torch at night, what's wrong?
Trend 4: Large model images have become a new volume point for smartphones
The wind and clouds are from the sky, and the dragon and snake rise from the land. The Year of the Dragon is destined to be an uneventful year for the smartphone industry. With the return of Huawei and the implementation of the device-side large model, various mobile phone manufacturers have begun to speed up their efforts to accelerate the R&D process and differentiation.
Among them, mobile phones have to roll up large models, and almost all mainstream domestic mobile phones have launched end-side large models. When the large model lands, the high probability of the volume is the play of images and images.
From the perspective of the current mobile phone side large model application, language ability, interactive update and image gameplay are the three main tracks. Among them, the most fresh and most enjoyable for consumers to experience the large model landing point is still in the image and video.
Therefore, it can be judged that the Year of the Dragon is most likely the first year of the large model image. From AI imaging, computational photography to large-scale model imaging, new mobile events are about to start.
In the distant future, it is a very noteworthy trend for AI models to change the logic of mobile phone interaction. A new species of terminal can already be seen at CES that is detached from touchscreen interaction. Although it is still very immature and takes time to develop. But it can be seen that the changes that native intelligence has brought to mobile phones have just begun.
Trend 5: Under the east wind of Apple, the domestic XR is replaced on the street
Recently, there was a joke that said that you went abroad to buy two Apple Vision Pro and came back to sell them. The money earned by the first unit can be reimbursed for machine wine, and the second unit can make a small profit, which is equivalent to free travel abroad and money to earn.
This reminds me of interviewing a VR manufacturer a few years ago. The person in charge told me that the team thinks that when Apple comes out with VR, the company's good days will come.
Okay, now that Apple is indeed out of VR, are the good days coming?
This problem will affect the next domestic VR and AR markets. In the case of Apple's banner and the Meta Quest 3 ecosystem becoming mature, the domestic XR brand should have ushered in an absolutely favorable development period. Just the concept of "apple replacement" can bring market feedback that cannot be underestimated.
In the next year, we should be able to see domestic AR glasses that have been relatively out of the circle to chase after the victory. The replacement of VR all-in-one machines has also begun to surface. The long-lost XR popularity seems to be really returning.
Trend 6: China's AI computing power is surging and surging at the forefront
Which type of tech company is the most motivated in the Year of the Dragon? It's not AI, it's not the Internet, the answer is IT companies.
In 2023, AI models will drive the explosion of IT needs for computing power, storage, enterprise networks, and data centers, and IT companies will usher in an epic development space. And there are magical "teammates" on the other side of the sea, who have been forbidden by Nvidia and can't help jumping sideways repeatedly, so they are mysterious.
By this year, there is basically no suspense about the iron curtain of AI computing power, and the ambiguous period between supply and demand has come to an end. Then the next thing is, ditch your illusions and get ready to fight.
Localization and AI computing power constitute the two biggest opportunities for IT companies. Moreover, after years of development, China's AI computing power itself has formed large-scale application capabilities and a relatively rich software and hardware ecosystem. At present, more than 10% of China's AI computing power has been deployed for localized substitution. This means that China's AI computing power has a foundation for development, and at the same time, there is also a huge market upside.
The two fulcrums of intelligence and localization are superimposed, and a rare outlet in the history of the industry is taking shape.
The story of domestic AI computing power also tells us that peace and joy are not waiting for it, but playing it.
Trend 7: The industry model, from wide to deep
The special feature of the AI model in China lies in China's abundant demand for industry intelligence and its huge potential for industry intelligence. We once said that China's big model is the big industrial model.
In the past year, major software vendors, AI vendors, and cloud computing vendors have explored the industry model market. They have even begun to advertise that MaaS is a new form of business that will replace SaaS.
It remains to be seen whether the MSSS model can be established. But it can be seen that the industry has a very strong demand for large models.
In the field of industry models, a layer of more profound changes will inevitably occur, that is, from a general-purpose model with a relatively simple model and a low degree of model fine-tuning, to a deep industry model oriented by key industries.
The inability of AI to penetrate into the core of industry production has always been a much-criticized problem. A large part of this is due to the fact that platform vendors are reluctant to over-invest in a certain industry, but want to achieve a one-to-many cooperation model to ensure that business profits are maximized.
However, this shallow model brings more user input and less output, and the doubts about AI technology are becoming more and more serious. Under the premise that the large model itself has great industry imagination, cloud computing and AI vendors will pay more attention to key industries rather than generalized model distribution in the future.
Among them, the financial model is the most noteworthy "pioneer industry".
At present, China's cloud computing industry itself tends to be conservative. Manufacturers generally reduce scale, reduce costs and increase efficiency. But retreating is not the way out. The solution of cloud computing in China is deep in the industry, and the Chinese market is different from the world.
Trend Eight: 55G is going global in the process of commercialization
In October 2023, during MBBF in Dubai, 13 operators around the world jointly released the first wave of 5G-A networks, marked 5G-A, or 55G is moving from technical validation to commercial deployment.
Subsequently, various 55G deployment tests are emerging one after another. A series of forward-looking, groundbreaking 55G testing has been successful across China.
In the next year of the dragon, we will certainly be able to see 55G is moving towards large-scale commercial use. There will be a full 5 on the market5G commercial solutions. Operators in China, South Korea, the Middle East, and some European countries will take the lead in launching 5Commercial procurement of 5G.
Huawei and ZTE represent Chinese communications equipment manufacturers, in 5The opportunity period of 5G is still leading.
Explore based on 55G's new industrial digital scenarios and terminal application scenarios are also 5The highlight of the first year of 5G commercialization.
Trend 9: Low-altitude economy, explore applications under favorable policies
People fly low, cargo goes high. These conceptual ideas of a low-altitude economy inadvertently became lively.
With the support of mobile network infrastructure, logistics system, drone technology, and autonomous driving technology, China has a large number of favorable factors for the development of low-altitude economy.
Therefore, in the context of exploring new digital formats, relevant policies for low-altitude economy have been introduced in various places. Although the industrial application of low-altitude economy is still difficult to form a scale in the short term. However, the next year should be a stage for various localities to carry out pilot exploration of application around low-altitude economic policies and give key support to relevant enterprises.
Unmanned aerial vehicles, unmanned logistics, agricultural digitalization, remote sensing, and even flying cars will all have better development opportunities than ever before.
Trend 10: Usher in grid-friendly and affordable photovoltaic power
If you go to the northwest for a road trip today, you will see the scenery of photovoltaic everywhere in the Gobi. At present, China's new and cumulative installed capacity of photovoltaic generators ranks first in the world. The total amount of photovoltaic power generation has exceeded 5% of the country's total power generation.
Under the "dual carbon" goal, photovoltaic power generation has two core missions. The first is to achieve grid friendliness and solve the problem that photovoltaic power generation is difficult to absorb and has a huge impact on the power grid. The second is to continue to reduce costs and complete the further parity of photovoltaic power generation.
With the support of a series of power electronics and intelligent technologies, these two goals are likely to usher in rapid advancement. For example, according to relevant data**, the power density of photovoltaic inverters will increase by 50% in the next five years, and through intelligent operation and maintenance and other technologies, the cost of many aspects of photovoltaic power generation can continue to decrease.
Perhaps, we ushered in affordable photovoltaic power, and even photovoltaic power as the main power is no longer far away.
Wrapping up
China's science and technology is just like the geography of this country, it is not a single landscape, but the collision and integration of countless landscapes. The changes in China's technology trends that can be clearly seen come from all sides. Among them, there are C-side and B-side, software and hardware. There is not only the proud posture of China's science and technology leading the world, but also the follower situation that China's science and technology still needs to imitate, learn and reproduce. There is a side of China-led exploration of the unknown, and there is also a side of last resort under the iron curtain of science and technology.
Facing up to the overlapping landscapes is a prerequisite for our discussions on Chinese technology.
In any case, "there is change" is the best change in itself.
Finally, if it is necessary to summarize and look forward to the changes in China's science and technology in the Year of the Dragon. Then maybe you can use two idioms about dragons, one is "let the dragon into the sea", and the other is "the cicada transforms into a dragon".
List of high-quality authors