After reviewing hundreds of AI hardware projects, we believe that this is not an era where parameters reign supreme.

In the upsurge of AI hardware investment, people always like to talk about technology, parameters, and ecosystems. But how can we see the essence and long - term value of a product? It may not be as closely related to these technologies and parameters as we think.

After reviewing hundreds of AI hardware products that have emerged in recent years, the Linear team has come up with a somewhat "counter - intuitive" view: what many teams need to learn most is not technology, but an insight into "human nature". The industry is used to imitating the functions of successful products, but it has always been difficult to replicate the naturalness of "going with your instincts" in those products.

Today's sharing comes from the recent observations and thoughts on AI consumer hardware startups by the investment team of Linear Capital. They shared their observations on the development trends in the AI consumer hardware field and their definition of a good product. Starting from the perspectives of "human nature, culture, and scenarios", they proposed a judgment model of "want to buy, easy to use, and willing to keep". Finally, they also looked forward to the path and direction of the globalization of Chinese AI hardware. We welcome more friends to communicate and discuss with us.

Based on our observations of the AI consumer hardware field this year, we can see that there are multiple development paths in this field, each of which may produce projects that exceed expectations and also has its own challenges.

However, we have a strong feeling that hardware provides the entry point, while software determines the imagination space.

Today's sharing will focus on "how to judge whether it is a good product" and share some of our recent thoughts.

Let's cut to the chase. What exactly is a good product? We have summarized a three - layer model, from the outside in: user scenarios → culture → human nature.

The outermost layer is user scenarios. The product must be viable in real - life moments. In the past, many Chinese consumer hardware products only focused on "stacking functions and parameters", while a truly good product is not about the functions or technology themselves, but about solving a problem or creating a unique experience in real user scenarios.

For example, the "Memories" function in the Apple Photos app plays background music during playback to help users quickly enter an immersive state of reminiscence when looking at old photos. When the user presses pause, prepares to speak or share, the system automatically lowers the background volume because it "understands" that you are likely to talk or show the photos to friends next. These are all designs that follow human nature in real scenarios.

Behind this is the designer's full understanding of what the user wants to do in this scenario. So when we look at early - stage projects, one of the core judgment dimensions is actually: Is the founder creating functions or creating a unique experience in a real user scenario?

The middle layer is culture. The product must be naturally accepted in different user environments. This is especially crucial for overseas - expansion teams. Apple FaceTime doesn't have a beauty - filter function, which reflects the sincerity and confidence of Western culture. In Eastern culture, beauty - filtering is, to some extent, a social etiquette. People need to confirm the relationship before deciding whether to show their faces. The same function is accepted differently in different cultures. Expanding hardware overseas is not about copying, but about creating an expression that can be naturally understood by local users.

The innermost layer is human nature, which is the foundation of the product. We can use dopamine and oxytocin as examples: Dopamine makes people constantly pursue new stimuli, while oxytocin consolidates these relationships to provide a sense of security in the present. A good product needs to find a balance between the two to last. A good product should not only make people addicted but also make them attached and keep using it continuously.

Ultimately, a good product starts from human nature, is nourished by culture, and is realized in user scenarios. These three points are all centered around the understanding of users, rather than simply starting from a certain technology and "looking for nails with a hammer".

For AI consumer hardware, we have summarized a standard testing flywheel: want to buy, easy to use, and willing to keep.

First, want to buy. Ideally, users should be able to understand "why I should buy it" within 30 seconds. Users don't need to understand the parameters; they just need to understand "what specific problem it solves" at first sight.

For example, Insta360 GO, in a highly homogeneous camera market, didn't talk about image quality, lenses, and parameters. Instead, it redefined the usage scenario - it is a "life recorder that is light enough to stick anywhere", which allows users to understand at first sight that it makes shooting easier and more part of daily life. We believe that most good products can make users excited with just one sentence or one video.

Second, easy to use. The user's first - time experience determines whether the product will have a second chance. There is a standard for judging whether a product is "easy to use": Users should be able to master it completely within 5 minutes without a manual or learning cost. Earphones that connect automatically when the case is opened are a typical example. You can use them right away without any operation. The reason why the 3D printer of TuoZhu can break into the mainstream market is that its design ensures that even beginners can use it right out of the box. Moreover, it has a MakerWorld community where users who don't know how to design can download model files for printing.

The problem with many failed hardware products is that the first - time experience is not smooth, the process is complicated, network configuration is difficult, and the prompts are unclear. If users still don't know how to use the product after 30 minutes, it basically loses the chance of being opened again. The more "light - weight and high - frequency" the AI hardware is, the more it needs to be extremely easy to use.

Third, willing to keep. What really determines the long - term value of a product is often the software capabilities. Hardware provides the entry point, while software determines whether users are willing to use it every day. This is especially obvious in AI consumer hardware. For software to make users stay, it must meet some underlying needs of users, such as:

The desire to share: Can the product make users willing to actively "show off their achievements"? For example, the recovery score calculated by the Whoop bracelet makes users naturally share after seeing the feedback, which is an embodiment of the user's "seeing" the value of the software.

A sense of achievement: Duolingo's consecutive check - ins are a typical example. By setting goals and providing various achievement awards, it satisfies the user's sense of achievement, making users naturally open the app at a fixed time every day without thinking.

In terms of quantitative indicators, we will pay more attention to weekly retention, monthly retention, frequency of active sharing, and number of active openings. These indicators together determine whether a product can truly become an AI hardware that is "used every day".

Finally, regarding the globalization of Chinese hardware, our basic judgment is that we don't necessarily need to copy the "soul" of the West.

European and American technology giants promote a "high - end, leading" value system, emphasizing high premiums and top - notch experiences. Chinese companies can take a different path, which we call "Inclusive Tech Innovation".

The core of this path is not to blindly pursue high - end, but to enable a wider range of people to enjoy good technologies that were once "luxuries" at an acceptable cost. If we can achieve this, what we create will be not only commercial value but also a new concept that can be accepted by more users around the world.

Of course, achieving this requires higher standards. It's no longer just about competing on parameters and cost. It requires us to understand both technology and products and truly understand the culture and human nature of different markets. Understanding the target market is a prerequisite, but the answers we come up with may be completely different from those of Silicon Valley companies. It is entirely possible to blaze a new trail.

Ultimately, both product development and investment are continuous practices. Theories can guide the direction, but the real path can only be forged step by step in the market. Practice is the sole criterion for testing truth and the only way to converge on the correct iterative direction of AI hardware.

This article is from the WeChat official account "Linear Capital". Author: Linear Capital. Republished by 36Kr with permission.