On April 18, 2024, Lenovo held the 2024 "Lenovo Tech World", and while releasing a number of new laptop products in one go, it also put forward the concept of "true AI PC" in a high-profile manner.
According to Lenovo, they believe that a real AI PC needs to have five characteristics, namely an agent with a built-in personal model that naturally interacts with users, local heterogeneous computing power (CPU/GPU/NPU), a personal knowledge base, an open artificial intelligence application ecosystem, and personal data and privacy security protection.
If you have paid attention to the "proposition" of AI PC in the field of consumer electronics before, you may feel that the five characteristics summarized by Lenovo have been very comprehensive and "great". It not only mentions the hardware level of AI PCs, but also takes into account the interactivity at the software level, user data security, scalable software ecosystem, and AI's self-learning ability.
However, as an old user who has been using Lenovo computers since the Y510 and Y550 eras, we may know more clearly about Lenovo's exploration process in the field of "AI PC" than many friends. Because of this, after seeing Lenovo's definition of "true AI PC", we have some questions.
Lenovo's AI PC road actually started as early as 7 years ago
When did Lenovo start "doing AI"? If you look back at the relevant historical data, you will find that in fact, as early as 2017, the theme of their Tech World was "filling the world with AI".
In that year, Lenovo proposed the "three elements" of AI of "algorithms, big data, and computing power", and has built its own AI innovation centers in many places around the world, and has also released a series of mass-produced products including 12-lead ECGs, AI service robots, and mobile phone artificial intelligence modules.
In 2018, we witnessed Lenovo's first "smart computer" product line.
According to the relevant information at the time, this batch of earliest Lenovo "smart computers" has the five characteristics of "always-on, remote management, privacy protection, voice interaction", and "AI computing power".
Although they are still a little different from today's products based on device-side large models in terms of concept, they have already run AI functions. And most importantly, the key to driving many of these localized AI features is not the well-known CPU, NPU, or GPU, but a "AI chip" developed by Lenovo.
Yes, as early as 2018, Lenovo had already thought of deploying low-power machine learning units into PCs to achieve some underlying AI features.
For example, on Lenovo's YOGA Book2, its ink-screen touch keyboard uses a self-developed embedded AI solution. By continuously learning the user's usage habits, it can automatically adjust the finger response parameters of the touch keyboard, so that the false touch rate of this EInk ink screen keyboard can be significantly reduced, and it can also identify the size of the user's touch area, so as to avoid the interference of the wrist and palm on the EInk touch keyboard.
Self-developed AI chips and AI functions are Lenovo's greatest confidence
That's right, in our Sanyi life's view, Lenovo's round of "true AI PC" announcement in 2024 can certainly make sense, but it somewhat ignores their previous long-term accumulation in AI, especially in the field of PC AI, and the persuasiveness has declined.
Especially when we consider the fact that Lenovo has produced its own AI chip and made it into PC products a few years ago, we are even more surprised that it is not advertised today.
From the machine learning chip that was first implemented in the YOGA Book2 and the model is unknown, to the Lenovo LA series controller that appeared for the first time in the YOGA Book 14S, and then to the AI chips such as LA2 and LA3 that are now widely used in Lenovo laptop products, Lenovo's self-developed AI chips have actually evolved "unconsciously" for many years.
In terms of function, today's LA series chips can realize functions including automatic camera perception (automatic screen lock when the user leaves the computer), automatic performance adjustment (automatic adjustment of different power modes according to the currently running software), WiFi automatic switching and other functions. On some Legion brand laptops, the high-end model chip can even recognize the current game frame rate and automatically control the performance output level, so as to save power and reduce heat generation as much as possible while satisfying the smoothness of the game.
And on the thin and light notebook that Lenovo is currently using, and Lenovo will launch in 2022, you can even see the online video super-resolution function based on AI algorithms. You know, even NVIDIA didn't get the video superresolution algorithm based on the independent display until 2023, and other brands of integrated display laptops don't have similar features so far.
In other words, this shows that Lenovo has the ability to develop its own AI chips on the one hand, and on the other hand, it does start earlier and do more in-depth than its competitors for the performance mining of "path" CPUs, GPUs and NPUs.
As a pioneer in AI PCs, Lenovo should be more confident
In fact, just a few days before the release of this content, we at Sanyi Life just criticized some mistakes and omissions in the current "AI PC" industry at the product promotion level at an AIGC theme summit.
In our opinion, the low-computing power NPU can't really become the practical core feature of the "AI PC", at least for the time being, and the simple voice assistant or AI drawing function is certainly fun, but it can't bring obvious help in real productivity or user experience for the time being.
Because of this, when Lenovo announced the concept of "true AI PC" in 2024, we are very clear on the one hand that Lenovo, as a technology pioneer in the domestic AI PC field, does have the corresponding qualifications and actual function accumulation. But on the other hand, the current promotion of Lenovo AI PC does not seem to seize its own rare advantages in the past, but is saying something similar to other brands, which inevitably seems a bit "cliché".
[The picture in this article comes from the Internet]