Before the world was opened, the computer industry was still in a chaotic era, and two Baidu stickers were born.
Tuba users focus on using funny hardware to match the best computers, and card bar users spare no effort to promote GDP development.
Now there are some new questions in the post bar: "Can your own graphics card be AI alchemy", "Grandpa Niu, take a look, 4060ti 16g or 4070 12g, which can run SD (Stable Diffusion)" and so on.
These questions are varied, but they all point to a common trend: AI PCs.
At present, there is no accurate definition to describe what an AI PC is, but in the "AI PC Industry White Paper" released by IDC some time ago, it is clearly written:
"Nearly six adults will consider AI as a factor when buying a computer."
Much of this craze comes from the unsparing hype of PC manufacturers, with Apple mentioning the Mac's AI capabilities at every press conference, and Microsoft even wanting to add an AI button to the keyboard.
But the most excited is a group of OEMs represented by Lenovo and Asus, after all, this wave of AI is an unprecedented good opportunity in their eyes-
They have suffered from Nvidia and Intel for a long time.
01
There is no autonomous OEM
Before talking about why AI PCs can make OEMs happy, we have to mention how traditional PCs are declining.
As of Q4 2023, global PC shipments have fallen for eight consecutive quarters, reaching 241.8 million units in 2023, down 14.8% from 2022, according to Gartner data.
This is the first time since 2006 (230 million units) that it has fallen below 250 million units.
The last wave of PC market recovery occurred around 2021, which may have been due to the demand for remote work brought about by the pandemic, but PC shipments continued to decline as the pandemic ended.
There are two reasons why people's desire to replace PCs is reduced: the lack of innovation of the PC itself and the replacement of PC use scenarios.
"It's getting boring, there's no need to change to a new one" is the common perception of PCs.
After years of development, the form of the PC has basically been determined, aside from the computer party (people who regard the computer as a work of art), the development of the PC today is more about the upgrading of hardware, better CPU, faster memory, and larger hard disks.
In this environment, the upgrade of each generation of PC actually depends entirely on the face of upstream hardware manufacturers, NVIDIA launched 4090, and the 3080 equipped with it can only be reduced to backward products.
The most intuitive example is the advertisement of a laptop, and the information of the CPU and graphics card is always the C bit.
But for most consumers, the perception of these hardware upgrades is not strong, no one will often change to a new computer for hardware performance, it took 5 seconds or 10 seconds to open PPT for ordinary people, the main one can be used.
The second reason for the shrinking PC market is that the functionality of PCs is being replaced.
With the improvement of the performance of mobile phones and tablets, we have fewer and fewer opportunities to use PCs.
If you want to watch TV series, you can use a smart TV with a better display effect; If you want to play games, you can choose a mobile game or a console like PS; If you want to read the news and read books, a lightweight tablet is more suitable.
The only thing we have to do is to get a job, but who wants to buy a new computer so often just to work? Anyway, when Silicon Jijun joined the company, the administration took a gadget purchased in 2017 and said to me:
This one also turns on.
Moreover, even these few use cases are still being stared at.
The iPad upgraded its system with an ad that "why should your next computer be a computer", Samsung launched Dex in an attempt to replace computers with mobile phones, and the former Hammer TNT desktop system.
In this context, the wave of artificial intelligence is coming.
02
An opportunity to turn over and be a master
Since the beginning of this year, a large number of advertisements for AI PCs have appeared on major platforms, among which OEMs have the largest voice.
The main reason why they hope that this wave of AI will change their lives is that the current definition of AI PC is extremely vague.
Microsoft detailed its definition of an AI PC in a report:
All AI PCs running Windows must be equipped with Copilot function keys, the ability to run Copilot natively, and a Neural Network Processing Unit (NPU) with a performance of up to 40TOPS.
It's a vague concept, and consumer perception is unusually weak.
In this way, OEMs have more than enough room to define AI PCs at the terminal/software level, taking advantage of the proximity to the consumer and capturing the minds of users.
This is also known as software-defined hardware.
Once the "xxx brand is AI PC" is formed, it will be able to take the initiative in the future market competition.
Specifically, in the AI PC industry, OEMs play a central role in integrating new AI technologies, providing heterogeneous computing power, developing AI operating systems, and driving technology convergence.
Through these key functions, they not only ensure the technical implementation of AI PCs and the optimization of user experience, but also promote the innovation and market growth of the entire industry, which is a key force to promote the intelligent transformation of personal computers and the popularization of AI technology.
So let's take a look, what tricks are OEMs playing on AI PCs now?
The first is Lenovo, which launched AI PCs earlier, and their AI is mainly based on Lenovo Xiaotian AI agents.
AI drawing, AI PPT, AI summarization, etc. can be carried out.
On Huawei's side, there are more AI functions.
Based on the Pangu model, Huawei's AI PC can quickly and accurately extract key points and generate summaries from massive amounts of audio, video, or real-time minutes, saving you a lot of time to grasp the key points of meetings and courses.
According to the press conference, Huawei's built-in AI space also features more than 100 agents, covering a variety of capabilities such as copywriting and programming.
ASUS is more interesting here, its AI function is divided into two parts, the +AI APP provided by Microsoft.
The first is the retrospective feature proposed by Microsoft Windows, which allows users to find what you have seen on the PC through the fragments of clues in memory, or real-time subtitles, etc.
Secondly, ASUS combines its advantages as an OEM manufacturer in hardware interoperability to make media management, AI sound enhancement and other functions.
In addition, AI PCs are based on powerful computing power and can run local large models, which are much stronger than cloud large models in terms of model response speed and data security.
The large model is a black-box application, and there is a high probability of data vulnerabilities. When ChatGPT first appeared, when it was asked to repeat a word multiple times, the model would deviate from the normal chat-style generation in some cases, and start to output text closer to the training data, which is simply an information leak.
emmmm...
Seeing this, will you have the same feeling as Sijijun: That's it? That's it, is this bit of AI really worth changing computers?
Most of the AI functions of AI PC seem to be completed by downloading AI software, and there are many AI plug-ins and AI tools that can be used like Doubao and Wenxin Yiyan.
The price of AI PCs is more expensive than ordinary computers, so why should consumers pay for it?
The advertising of the AI PC is well done, and the vision portrayed is fascinating.
However, OEMs can do more than just now (squeeze toothpaste), and the advantages of OEM terminal integration and AI operating system mentioned above are not clearly reflected.
OEMs should still take advantage of their advantages as direct contact with consumers, and work harder at the system level and in the AI ecosystem.
Imagine if the AI assistant that comes with the computer can directly modify the BIOS, wouldn't it be more distinctive?
The AI PC market, worth hundreds of billions, has brought new opportunities to OEMs, but it is still necessary to grasp it.
Don't try to make consumers feel like AI PCs are just a gimmick.
End of full text.