Produced by | Bullet Finance
Author | Uncle Yu
Edit | lightning
American Editor | Qianqian
Audit | Ode
In 2024, a new wave of technology will sweep the world. At the beginning of the year, the deep integration of hardware, artificial intelligence (AI) and large models has become the most eye-catching focus, which marks the full opening of the AI era.
However, in this technological frenzy, many enterprises are facing a common problem: although large-scale model technology emerges in an endless stream, the practical application is still on the surface, and it is difficult to truly integrate into enterprise operations, let alone promote industrial upgrading. In this context, the industry urgently needs a truly effective roadmap to help enterprises quickly build an AI ecosystem.
Now that AI industrialization has reached a new stage, it is necessary for everyone to "benefit from the development of artificial intelligence, so as to create a sustainable and beautiful future that 'fills the world with AI'." ”
Since the release of Open AI in October last year, especially with the launch of the GPTS store, AI Agent has gradually become recognized by Internet giants as the best entry point in the roadmap for the integration of AI and large models by enterprises.
At the just-concluded 2024 Lenovo Innovation and Technology Conference, Lenovo's timely upgraded "full-stack AI" strategy, the most important of which is the "integrated and multi-terminal" strategy, especially the "Lenovo Xiaotian" personal agent that landed on AI PCs, is considered to be a milestone product leading industry change. It pushes the application of AI and large models in the enterprise to a new stage of development.
As a new type of personal agent, "Lenovo Xiaotian" can not only understand the user's natural language instructions, but also actively learn and adapt to the user's needs and habits, so as to provide users with more personalized and intelligent services. For enterprises, this means that they can use "Lenovo Xiaotian" to extend the large model technology to AI PAD, AI mobile phones and other smart devices through the ultimate carrier of AI PC, and finally apply it to actual scenarios to realize the commercialization of the technology.
1. Personal intelligence becomes the core of the landing
With the continuous breakthrough of artificial intelligence technology, large models have become the new favorite in the field of science and technology. However, for many enterprise users with limited technical strength and unfamiliar with large-scale model training and optimization, it is still a challenge to navigate this complex technology.
The training and tuning of large models is complex and time-consuming, and the technical threshold is high, and even if the training is successful, how to maximize its performance in practical applications is a major problem. In addition, the task specificity of the large model makes it incapable of dealing with diverse enterprise needs.
However, at the beginning of 2024, a technological change led by Lenovo's personal agent "Lenovo Xiaotian" is quietly changing this situation.
As a bridge between users and large models, the unique advantages of personal agents are rapidly becoming apparent. Traditional application interactions are gradually being replaced by intent interactions, and users experience unprecedented convenience and efficiency when using large models. At the same time, the intelligent processing capabilities of personal agents represented by "Lenovo Xiaotian" provide users with strong support when completing tasks.
In particular, the processing of personal information and knowledge, which is the most extensive demand, with the support of "Lenovo Xiaotian", only needs to describe the requirements to it through natural language, and the personal agent system can intelligently decompose the task and call the corresponding plug-in, and use the personal large model tested by the computer to efficiently complete tasks such as sorting out relevant information.
For example, in Antarctic scientific expeditions, scientists are generally faced with scientific research problems such as communication and manpower constraints, and it is not easy to efficiently complete the collection and processing of scientific research information in such an environment. Now, with the help of AI PC personal agents, scientists can rest assured that they can use the computer-matched personal large model to quickly process the collected data and information, and can generate various types of answers according to their own needs, greatly saving scientists' time. Even, for many papers on the Internet, AI PC personal agents also intimately translate and refine the content, so that scientists can grasp the required information in the shortest time, greatly improving the efficiency of scientific research.
Yu Hai, Senior Vice President of Lenovo Group, General Manager of Commercial Product Center and Global SME Business, said in his speech at the 2024 Lenovo Innovation and Technology Conference: "The era of AI PCs has arrived, and PCs will evolve from personal computers to artificial intelligence personalized computers. The magic of turning a computer from a tool into a partner that understands you comes from an individual agent. ”
In fact, in this new era of AI, the personal agent system is playing a role in introducing AI into personal life, just as the role of Baidu in the PC era and WeChat in the mobile Internet era is as important.
Tech giants have already sniffed this huge opportunity and have laid out one after another.
Voyager, an AI agent launched by a well-known American chip giant, has shown amazing strength after connecting to GPT-4 and is unique in the game "Minecraft", while another Internet giant has teamed up with American universities to create an interactive sandbox environment called "Westworld", in which generative AI agents can simulate human behavior and live freely in the virtual world.
It is worth noting that although these personal agent systems have made remarkable achievements at the Internet level, there is still a gap in the entrance to truly realize the integration of software and hardware.
This means that in the future science and technology competition, whoever can take the lead in launching a personal agent carrier that integrates software and hardware is likely to become the leader of the new era of AI.
2. AI PC is the preferred carrier
Most of the existing large models and individual agents are still limited to the cloud, which is difficult to meet the personalized needs of users. At the same time, the actual experience of individual agents that rely on cloud services tends to lag as the number of users increases. These problems have limited the popularization and application of AI technology.
In order to break this situation, it has become imperative to find the ultimate carrier that can carry AI technology. In this regard, Lenovo Group was the first in the industry to release the AI PC, and pushed it to the forefront of the technological wave as a new terminal device.
Yu Hai praised AI PC as a "new service carrier", emphasizing that it is locally embedded in the personal agent "Lenovo Xiaotian" and linked to the Tianxi AS system to realize core features such as natural language interaction and running personal large models. This not only solves the lag problem caused by relying on cloud services, but also provides users with a smoother and more intelligent interactive experience. At the same time, the integrated hybrid AI computing power and open AI application platform of AI PC provide a reliable guarantee for user data security and personal privacy.
Yang Yuanqing, chairman and CEO of Lenovo Group, said at the summit that AI PC products should have five characteristics: intelligent twins embedded with large personal models and natural interaction with users, personal knowledge bases, local heterogeneous AI computing power (CPU/GPU/NPU), open artificial intelligence application ecology, personal data and privacy security protection. These characteristics not only define the future development direction of AI PCs, but also fully demonstrate Lenovo's innovation and leadership in the field of AI PCs.
Among them, the intelligent twins and personal knowledge base embedded with the personal large model and the user's natural interaction make the AI PC closer to the user's needs, the standard local heterogeneous AI computing power (CPU/GPU/NPU) ensures efficient and intelligent task processing, and the open AI application ecology, personal data and privacy protection solution provides users with a safer and more convenient use environment.
In the AI PC defined by Lenovo, the personal agent is crucial, as it is a new bridge between the large model, the terminal and the user. Driven by a large model embedded in the local area, the personal agent can realize natural interaction with the user, shuttle through the AI terminal, bring users a new AI experience in multiple scenarios such as work, study, and life, and undertake the tasks of the user's personal assistant.
The rise of AI PCs is the result of the combined effect of generative AI technology characteristics and the natural attributes of PCs.
From a technical point of view, the multi-modal natural language interaction capabilities and compressed general scene service capabilities of AI large model technology are perfectly matched with the status of PC as the strongest personal computing platform. In addition, as users pay more attention to data security and personal privacy, it has become an inevitable trend to implement AI large models locally, and PCs, as the most trusted terminals, are undoubtedly the best choice.
In terms of specific capabilities, AI PCs fully release the relevant capabilities of individual agents by directly deploying computing power offline and using the joint computing of NPU and CPU. With the hybrid invocation of individual large models and general large models, the ability of individual agents to meet user requirements has been significantly improved. This makes AI PCs more efficient and intelligent when handling complex tasks, bringing users an unprecedented convenient experience.
Therefore, the rise of AI PC will create a new era, just as smartphones lead the era of mobile Internet and cloud services bring high-speed interconnection, AI PC is expected to become the ultimate carrier of the AI era. IDC predicts that the assembly ratio of AI PCs in China's PC market will rise rapidly in the next few years and reach 85% by 2027. In the "AI PC Industry (China) White Paper" jointly released by Lenovo Group and IDC, it is also emphasized that 2024 will be the first year of AI PC.
In this process, Lenovo, as a pioneer and leader of AI PC, has made remarkable achievements.
On the one hand, Lenovo actively promotes the capability development of personal agent systems and is the first to release AI PC products in China, and on the other hand, Lenovo works with ecosystem partners to create a new pricing system under intelligent computing to balance the relationship between intelligence and price. In addition, Lenovo is also committed to building a platform for natural interaction, promoting the construction of application ecology, and accelerating the development of other intelligent terminal products.
At the recently concluded Lenovo Innovation and Technology Conference, Lenovo officially released the AI Yuanqi Edition series and a number of new smart terminal products, fully demonstrating its positive progress in the formation of the AI PC industry ecology.
On the same day, the AI Terminal Industry Ecological Innovation Forum was held in the Shanghai World Lounge. At the forum, Lenovo and the China Academy of Information and Communications Technology (CAICT) jointly initiated and prepared the industry's first AI terminal industry ecological alliance with leading enterprises in the industry.
As a representative of the sponsor of the preparation, Amu, vice president of Lenovo Group and chief strategy officer of Lenovo China, said at the scene that the AI terminal industry ecological alliance will be committed to promoting multi-level technological innovation collaboration, jointly promoting a standardized adaptation certification system, lowering the threshold for AI application innovation, jointly promoting the prosperity of the AI application ecosystem, and helping to develop an open, win-win and sustainable Chinese AI terminal industry ecology. Previously, Lenovo has launched the "AI PC Pioneer Action" for developers and users, and launched a number of programs and competitions aimed at stimulating innovation and promoting the development of the industry.
This is actually an important measure led by Lenovo to promote the landing of AI terminals in China.
3. Why Lenovo
In the field of AI PCs, Lenovo has become the industry leader and wind vane. The formation of this status is not accidental.
First of all, Lenovo has provided unique conditions for the implementation of new technologies with its comprehensive and powerful "end-side" layout - the first global PC market share, the top three tablets in the world, and the seventh in the world's mobile phones.
Based on a deep understanding of traditional PCs and a keen insight into the development trend of large models, Lenovo innovatively proposed the strategy of "one body and multiple terminals". Among them, "all-in-one" refers to Tianxi AS (personal intelligent body system), which will be loaded into each of Lenovo's intelligent devices, to achieve natural interaction between customers and devices, intention understanding, application and personal knowledge base call and other skills, and can be interconnected across devices to complete the distribution and relay of cross-device tasks;
At the just-concluded Lenovo Innovation and Technology Conference, Liu Jun, Executive Vice President of Lenovo Group and President of China, introduced Tianxi AS (Personal Intelligent Twins System), which promoted the commercial implementation of personal intelligent twins to a new height. According to Liu Jun's introduction, Tianxi AS includes a personal agent driven by a local large model - "Lenovo Xiaotian", a personal knowledge base, and an AI application platform, supplemented by Smart Connect cross-device connection and data security and privacy protection capabilities.
Second, the rise of AI PCs is inseparable from the support of hybrid large models. Because enterprises often put forward higher requirements for data privacy and the effectiveness of solutions. The hybrid collaboration of the general large model deployed on the cloud side and the customized large model deployed on the device side has become the most suitable solution.
Lenovo is also at the forefront of this regard, providing enterprises with the best choice to protect data privacy and ensure the effectiveness of their solutions through the synergy between cloud and device. At the same time, Lenovo also precipitates long-term customer service experience and technology into methodologies and service solutions, and provides full-cycle service support including AI platform tools and intelligent hybrid cloud platforms to ensure the stability and efficiency of services.
Finally, behind the AI PC layout, it is Lenovo's powerful full-stack AI layout that provides strong support. Since 2017, Lenovo has started a full-stack AI layout based on the new IT architecture of "device-edge-cloud-network-intelligence", covering the whole field of AI-embedded intelligent terminals, AI-oriented infrastructure and AI-native solution services, which has become the most solid foundation for the promotion of AI PCs. At present, Lenovo is the first AI solution and service provider in China to complete the full-stack AI layout.
From the perspective of technical characteristics, AI PCs rely on the large model technology of individual agents to invoke them, and finally can carry out multi-modal natural language interaction, and still have the ability to serve general scenarios after compression, which requires strong AI computing power for inference, and needs to fine-tune and personalize services based on personal data and privacy information.
The heterogeneous intelligent computing platform it builds can match the optimal computing infrastructure according to customer needs, and pool and optimize the overall computing power, so as to meet the comprehensive needs of customers for the overall computing power platform and provide a strong guarantee for Lenovo to achieve hybrid AI deployment.
Crucially, at the device level, Lenovo's product line is particularly comprehensive, ranging from general-purpose devices to high-performance computing products such as AI servers and AI storage. Especially in the field of AI servers, Lenovo has fully demonstrated its leading strength in AI computing power with its complete AI server product matrix and excellent performance improvement capabilities—the performance of large model training servers has been increased by 9 times, and the inference performance has been increased by 30 times.
Currently, Lenovo is one of the world's top three AI infrastructure providers and the world's No. 1 provider of scientific computing.
Therefore, Lenovo can meet the needs of different users for the implementation of hybrid large models in the AI era, so as to provide optimal solutions to customers' real business pain points, so that large model technology can be truly implemented in customers and industries.
Amu said in a previous interview that AI is actually like electricity-like infrastructure, which will penetrate into every terminal silently, and the full-stack AI tools provided by Lenovo will become a partner in everyone's work and life through the birth of artificial intelligence, which is generating a huge commercial market.
As a leading player in the PC industry, Lenovo has been at the forefront of change over the past four decades. From launching PC products that revolutionize the user experience to firmly occupying the leading position in the market, Lenovo continues to lead every generational upgrade of PCs. Now, with the widespread application of generative AI and large models, its AI PC is once again leading the industry to change dramatically.
In this context, Lenovo is expected to be higher. In this regard, Yang Yuanqing said: "Hybrid artificial intelligence is the representative of new quality productivity, which will greatly accelerate the intelligent transformation of all walks of life. The science and technology industry must seize this once-in-a-century growth opportunity, establish ecological alliances, form industry synergies, open innovation, and seek common development. ”