Time has proven that big models cannot save the consumer electronics market.

Over the past year, the combination of big models and terminal hardware has undoubtedly been the darling of the technology industry. On the one hand, as a software technology, AI big models aim to enter daily life and be widely used by the public, and hardware is a necessary way to lower the threshold. On the other hand, in 2023, the consumer electronics market performed poorly overall, hit by tight consumer budgets and weak demand.

In this way, the integration of AI big models and consumer electronics across all fields can be described as a natural and mutually beneficial process.

Can AI technology centered on big models lead a new round of hardware innovation cycles? Today, it seems that "the future has arrived, but unevenly distributed."

Big models remain the mainstay of traditional mass markets such as smartphones and laptops, but when we broaden our perspective beyond mainstream categories, the situation is far from optimistic.

Whether it's concept products vigorously promoted by industry giants like Zuckerberg and Google, such as Ray-Ban Metas smart glasses, neural wristbands, Google Glass with Gemini, and Samsung Galaxy Ring, or seemingly interesting tech innovations like AI Pin, Rabbit R1, and Limitless Pendant (AI+wearable), they all share an unreliable aura of familiarity.

Today, it seems that big models are not the universal key to the consumer electronics market.

AI will first enter our lives through mainstream hardware, reviving traditional dominant categories that have been under pressure.

Most other AI hardware that tries to "get in on the action" will end up just "getting nowhere." Let's use a "loneliness index" to understand which AI hardware is destined to disappoint.

Among the wave of AI terminal innovation, one of the most vocal categories is new hardware forms born from big model capabilities, such as the popularity of AI Pin and Rabbit R1, which stem from their novel concepts beyond traditional smartphones.

At the MWC conference, startup Humane shouted the slogan, "If a phone can do it, AI Pin can do it too," allowing users to interact through voice dialogue and laser display. Previously, Time magazine named it one of the best inventions of 2023.

Closely following was Rabbit R1, which claims to be equipped with the advanced AI big model LAM, enabling complex data processing and learning tasks. It has a display screen, providing a more intuitive interaction experience than traditional smartphones.

Consumers tired of traditional smartphones are naturally full of expectations for these new hardware forms.

However, as products enter reality and users gain in-depth experience, problems gradually emerge:

Broken promises. The AI functions promised by AI Pin and Rabbit R1 in their promotions were not fully realized, with many features resembling castles in the air rather than the revolutionary technologies they claimed to be.

Disappointing experience. These two AI hardware devices had numerous issues in actual use, such as latency due to insufficient computing power, system instability, and lack of high-frequency functions, leading to poor user experiences.

Market cooling. As returns increased and negative reviews emerged, market enthusiasm for these two hardware devices rapidly cooled, and the bubble burst quickly. Sales data disclosed by tech website The Verge showed that from May to August this year, AI Pin had more returns than purchases.

Ultimately, these emerging hardware devices have captured some vague needs (pain points in traditional smartphone interactions), but the product solutions offered are immature, making it difficult for consumers to develop deep experiences and product dependence, hindering the formation of real product value and long-term commercial success.

Facing multiple disappointments from the public, media, and investors, the initial enthusiasm will turn into a catalyst for accelerated value "evaporation."

An academic saying goes, "Monkeys building ladders may seem to improve with each attempt, but they'll never truly reach the moon." This aptly describes AI wearable devices.

Big models have breathed new life into terminal devices, and a batch of wearable devices are attempting to recover from the wreckage of the previous cycle, including:

Smart glasses. Meta is collaborating with Ray-Ban to create new smart glasses that use multimodal AI. You just need to say, "Hey, Meta, look and tell me what this is." They identify what you're looking at and answer related questions. Zuckerberg also mentioned that the most advanced smart glasses can display full-field vision, allowing future interactions with holograms appearing before your eyes. Meta is preparing to showcase prototypes of holographic glasses to the public but doesn't plan to sell them on a large scale.

Smart wristbands. Similar to fitness trackers, they can sense subcutaneous neural signals, enabling typing, controlling home devices, sending messages, and operating computers through subtle hand movements or even mental commands. However, second- or third-generation products are needed for maturity.

Smart rings. Samsung's Galaxy Ring enhances user interaction through AI functions, earning praise from the media as "the next-generation AI hardware."

Smart jewelry. LimitlessAl introduced the wearable pendant Limitless Pendant, claimed to be the world's smallest AI wearable device, offering all-day recording and dialogue memory summarization functions. It simplifies APP usage through dialogue and sold over 20,000 units within five days of its launch.

It's not hard to notice that similar products emerged during the peak of smart consumer electronics from 2014 to 2015, producing many benchmark products. Examples include Google Glass, Vuzix M100 smart glasses for enterprises and developers, Nest Weave smart pendant for tracking health and activity data, and smart bands and watches from various brands. Many of these products have been discontinued or replaced by subsequent versions.

In the era of big models, some inherent flaws in wearable devices have been alleviated to some extent. For instance, big models provide more powerful data processing and personalized interaction capabilities, enhancing user experience to some degree. With advancements in computing power from advanced chips, some tasks previously processed in the cloud can now be completed locally.

So, can they become the optimal form of AI hardware? Clearly, fundamental issues persist.

The primary constraints preventing widespread adoption of wearable devices are not technical algorithms but functionality (whether they solve previously unsolvable problems) and ecosystem (whether they offer diverse service options and commercial barriers).

Currently, many AI devices still lack a clear market positioning. For instance, AI pendants without speakers or screens raise doubts about their practicality. The supply chain is yet to mature, lacking production, application, and third-party developer ecosystem support, affecting devices' cost-performance ratio and function expansion.

Therefore, big models for wearable devices are like monkeys with new ladders—seemingly improved capabilities but unable to genuinely impress users or achieve commercial success.

Of course, for consumers, it's not entirely a "waste of money." After all, these big model-integrated hardware devices retain some functionality in their original forms. At least, wearing them makes others immediately recognize you as a "tech enthusiast," offering ample emotional value.

Are there any relatively reliable AI hardware options? Yes, there are.

These products share several characteristics:

1. Mature supply chain system. Software-wise, they can utilize leading foundational general big models to meet basic commercial demands. Hardware-wise, they feature mature forms and efficient supply chain development. Examples include Plaud Note recorders, AI mouse recorders, Curio's AI toys, and TWS AI earbuds. The "+AI" model built upon existing product functions often ensures development efficiency and product quality.

2. Clear business model. These mature product forms are easily imitated and replicated. To avoid fierce competition in a red ocean, high-barrier business models must be established. Take AI stuffed toys as an example. Children don't demand high AI illusion technology, but parents, as consumers, pay close attention to safety (materials, durability, interactive content), IP (renowned IPs, designs), etc. Startups can more easily establish competitive barriers in areas like refined interactive content algorithms, safety and controllability, educational knowledge and functions.

3. Avoiding competition with large companies. Large tech companies possess robust R&D and marketing resources, making it challenging for startups to compete. Innovating outside their scope, such as multimodal big models + OTC hearing aids, can yield growth rates exceeding general Bluetooth earbuds. Moreover, hearing aids for mild hearing loss in middle-aged individuals represent a niche market that large companies seldom enter or are unwilling to pursue.

Currently, vertically segmented fields with mature hardware forms are more likely to benefit from big model integration, offering the greatest potential for commercial success in AI hardware. Unfortunately, these "small but beautiful" products often have limited market sizes, unable to keep pace with mainstream categories like smartphones and smart computers, limiting their impact on the overall recovery of the consumer electronics market.

Consumer electronics hardware follows a "12-year rule." Generally, due to advancements in communication network technology and computing power, 5-10 years bring about significant technological upgrade cycles. Today's ubiquitous smartphones are over a decade old. From this perspective, the surge in AI hardware popularity stems from a combination of technology-driven advancements, user expectations, and industry cycles.

The ultimate purpose of consumer electronics lies in "interaction"—intimate conversations between humans and machines, silent collaborations among machines. The emergence of big models has revolutionized interactions, rejuvenating the entire consumer electronics industry.

Over the past year, we've witnessed the flourishing of AI hardware and experienced market chaos.

AI hardware dominated by big models is in an initial stage of vibrant creativity and intense transformation, transitioning from wild growth to stability and unity. At least, the consumer electronics market is no longer silent.