A new cycle from a broad historical perspective: Can generative AI make PC great again?

Since its birth in the 1970s, PC (personal computer) has reached the age of "knowing destiny". By borrowing Huang Renyu's "Big Historical View" and Kondratieff's "Kangbo Cycle" theory, and examining the origins of PC from a broader perspective, we may be able to see what its future will look like.

PC plays an important role in the IT (information technology) field, and its fate is closely related to the development of the IT industry. The PC originated from the Turing machine, information theory and cybernetics in the middle of the last century. It promoted the rapid development of information technology and became the most pioneering product in the 1980s and 1990s. It reached its peak in the Internet wave after 2000. However, after picking the "low-hanging fruit", the IT industry entered an innovation bottleneck period, and PCs began to gradually decline

Artificial intelligence started almost at the same time as information technology, but due to various reasons it failed to achieve a smooth development curve and progress was slow. However, in the past 5 to 10 years, the technological innovation and marketization process of artificial intelligence have accelerated significantly, especially the wave of generative artificial intelligence sweeping the world this year, which has made people see the dawn of the generalization of artificial intelligence. . This means that on the road to intelligence, personal computers will make small and fast progress, and even start a new acceleration cycle

Data and forecasts from third-party research institutions confirm that the PC market may usher in a new cycle. According to the latest report released by IDC Consulting, global PC shipments have increased in the past two quarters, and the year-on-year decline has slowed down, indicating that the market has gradually bottomed out. More importantly, "Generative AI is expected to become a watershed in the PC industry. Although there is currently no clear practical application plan, the development prospects of AI PC are promising"

John Roese’s outlook on the development trends of the technology industry in 2024 and beyond

PC giants have begun to take action and express their support for the future development of artificial intelligence, from a market perspective. At the recent Dell Technology Group 2024 Technology Forecast Media Communication Conference, the company’s global chief technology officer John Roese said: “Artificial intelligence is the center of the development of the future world, and it can be applied to production through edge technology and rely on zero trust guarantees. Security, and ultimately obtain continuous power from quantum technology to achieve the performance and efficiency required by global systems." In the field of client artificial intelligence, Dell Technologies has become the industry leader, and its development trajectory is worth learning from other companies

Obviously, the pace of generative AI from theory to practice is accelerating, and the implementation path of AI PC is becoming increasingly clear. In the new cycle, can PC be great again?

Client AI is expected to become the biggest "dark horse" in the generative AI track

When it comes to generative AI, the first scenario that people think of is usually this: large models are trained and inferred based on super clusters on cloud platforms, which consumes huge computing power and may also bring about revolutionary breakthroughs.

In fact, the evolution of generative AI does not only stay in the cloud, but is expanding from the cloud to the edge. It will also exist in users' future laptops, desktops, workstations and other personal computing terminal devices.

This new paradigm is called client AI (on-client AI), which can deploy and execute AI processes directly on the user's local computing device, thereby bringing massive processing power closer to the end user, so that the user no longer You need to rely on cloud computing infrastructure to obtain powerful AI capabilities

Compared with cloud AI, client AI can improve security and efficiency while reducing data costs: In terms of privacy and security, local processing means users do not have to worry about personally identifiable information and intellectual property being affected by potential threats in the cloud. ; In terms of response speed, from organizing files to enhancing camera functions, client AI can quickly produce results without affecting the device's processing capabilities; in terms of cost control, local processing can reduce data transmission requirements and significantly reduce the use of cloud computing resources. the costing payment

John Roese, Global Chief Technology Officer of Dell Technologies

The rewritten content is: In this sense, client AI is likely to become the most potential competitor on the generative AI track and stand out in the process of realizing dreams. As John Ross puts it: “Enterprise focus on generative AI will shift from broad experimentation to top-down strategic planning and select a few truly revolutionary directions. By 2024, we will see The first generative AI enterprise projects mature."

If the big tree of client AI has the opportunity to bear "fruits of midsummer", then how will its specific growth path be interpreted? Who will play the role of pioneer?

Multiple parties work together to explore the evolution path of client AI

Before the rise of generative AI, client AI had already begun to emerge. A series of attempts represented by PC intelligence have greatly improved the productivity level of terminal equipment, and also significantly improved the user's application experience

Dell Intelligent Tuning Software (Dell Optimizer) plays an important role in OptiPlex business desktops, Latitude series notebooks and other devices. It is an excellent software. Dell Optimizer uses artificial intelligence to automate configuration, truly turning computers into users’ life assistants and work tools by accelerating application running speeds, extending battery life, and blocking unnecessary background noise during video conferencing

Faced with the severe test brought by generative AI in terms of computing power and other dimensions, the workstation, as the performance leader among PCs, was the first to respond. Dell Precision workstation supports multi-GPU acceleration and CPU multi-threading, helping application developers to train large AI models with 5 million to 500 million parameters locally on the device anytime, anywhere, for model preparation, development, prototyping and deployment of generative AI. Protect and protect.

Of course, there is still a long way to go before we can run large-scale language models on PC devices and use them as a basis to fully empower artificial intelligence applications. Currently, one of the most popular techniques is Model Distillation, which trains smaller models to mimic the behavior of larger models. These smaller models can be deployed on users’ end devices, complementing larger models in the cloud

"Model distillation" is just the tip of the iceberg. In the context of the popularity of generative artificial intelligence, the language modeling, machine learning and other capabilities of client devices such as PCs are facing a window period for leapfrog growth, and systematic and even disruptive technological innovations are about to emerge. John Ross believes that by 2024 to 2026, there will be more artificial intelligence PCs based on CPU, GPU, and NPU architectures. Generative artificial intelligence applications will also have higher demand for distributed processing capabilities, and client artificial intelligence will Intelligence has considerable room for growth

The era of generative AI has just begun. Diversified and multi-level cooperation and collaboration are the necessary ways to make greater progress in client AI. There will be more cooperation models in the future, and client AI will achieve fruitful results

The ecosystem of new generation clients accelerates evolution

It can be said without a doubt that a new generation of client platforms and terminal ecosystems are rapidly taking shape. At the same time, generative artificial intelligence capabilities distributed at the edge and the cloud will also evolve together. It is foreseeable that a new cycle of PC has arrived

The future client platform will have many new applications, which will provide strong support for remote collaboration meetings and intelligent security solutions. Personal computing in the future will include adaptive user interfaces, automatic support and other generative AI experiences, and the integration of cloud and local computing will be more conducive to AI-driven personalization, system optimization and predictive operations, thus bringing about comprehensive User experience improvement

As far as the PC field is concerned, its role will evolve from a productivity tool to "driving overall human performance." In the future, the main method of command and control interaction will no longer be typing on the keyboard. There will be more non-text methods to achieve two-way interaction between humans and PCs - based on the user's voice, visual instructions or gestures, the PC can interpret their emotions and Communicate changes in thinking and deliver expected results in a way that best satisfies users. At the same time, the way users use PCs will change from search to prompts, from reading to understanding, and from editing to guidance.

From a broader perspective, generative artificial intelligence is not the only force driving client changes. The development of other core technologies will also have an impact on the entire ecosystem. John Ross believes that the establishment of multi-cloud edge platforms, zero-trust technology, standardization and certification are crucial, and quantum computing will bring a huge leap forward for artificial intelligence systems. He said, "Hybrid quantum systems are expected to become the computing basis of modern artificial intelligence, and the work of artificial intelligence will be dispersed across a series of different computing architectures, including quantum processing units."

The changes in the new generation of clients are exciting, and the stories in the new PC cycle are worth looking forward to. On the long journey full of uncertainties, Dell Technologies will focus on intelligence, security and becoming a trustworthy end-to-end AI consultant, continue to increase investment, and work with industry partners to build a beautiful picture of the new digital world.

The above is the detailed content of A new cycle from a broad historical perspective: Can generative AI make PC great again?. For more information, please follow other related articles on the PHP Chinese website!