Produced by |Tiger Sniff Technology Group.
Author |Maruto Mountain.
Edit |Wang Yipeng.
Header |Tiger sniff shooting.
In the marketing industry, there is a classic interview question: sell me this ballpoint pen for xx's **.
But if the question comes from an interviewer at an AI hardware company,Maybe just try to answer, "It's AI-driven." ”
While this may sound absurd, that's exactly what some consumer electronics companies are doing today.
These vendors are trying their best to show the world that artificial intelligence can be deeply integrated with hardware in various scenarios, but most of these products are of greatest significanceIt may be to verify how to define a product through "brain supplement" when the demand does not exist.
Of course, similar stories have happened many times in the history of consumer electronics, and new technologies are often accompanied by a large number of experimental products in the early stage of landing and exploration.
However, compared with the technology pioneers of the past, many manufacturers who are committed to AI hardware today are very aware that their products are not even "conceptual".
AI equals brand upgrading?
An interesting fact is that in terms of the boundary expansion of AI hardware, traditional home appliance companies may be the most radical in the industry.
Two years ago, Samsung showed off a spherical robot called "Ballie", whose biggest selling point is its ability to follow the user's movements and complete some simple interactions in the home scene. Maybe Samsung also realized how bad this idea was, and after Ballie debuted that year, it disappeared.
But not long ago, Samsung showed this product again after completing the tinkering of the ballie. According to Samsung's official statement, the upgraded ballie is positioned as an "AI companion robot", with its own projection function, which can be used as a user at any time, and can also be linked and controlled by smart devices at home.
***samsung electronics
Throw away the tall packaging concept and the tools to control smart devices, we generally call them gateways. The question is, who would need a gateway and projector that can follow them around?
LG, also a Korean company, also showcased a similar companion robot product and defined it as an "AI butler" to help users patrol their homes and observe the condition of their pets. LG also specifically mentioned that it can analyze the user's mental state by capturing the user's expression and voice, and if the user is in a bad mood, it will also be uplifted.
But it is still that problem, except for a few specific scenarios such as early education, even large language models that have been fully trained and adjusted parameters do not seem to be enough to support the robot to understand the user's semantics in real time and accurately, so what is the demand basis for companion robots?
Dipanjan Chatterjee, Principal Analyst at Forrest, made an interesting point on his personal social **:No one wants to show up at a drag ball in plain clothes.
The implication is that any company must cater to the mainstream of the industry and show its own actions in AI.
But the author believes that what can really make AI land on hardware is precisely those unpretentious "casual clothes".
Also take Samsung and LG as an example,This year, both of these have launchedAIContif for TV,And the focus is not on the gimmicks of human-computer interaction,Instead, focus on improving the picture quality of TV through AI chips and neural network models。
To put it simply,With the blessing of AI chips,TVs from these manufacturers can now analyze each frame of the picture,And intelligently supplement the details of the picture,For example, denoise the background,Sharpen the edge of the main body of the picture,Make the picture clearer。
Although it is not clear whether it can achieve the "4K to 8K" effect advertised by the manufacturer, but this technical path does solve the user's hard demand for TV:In the case of low quality of the film source,Improve the picture quality experience through hardware。
Perhaps this technology sounds far less than "multimodal models" and "generative AI"**, but at least for hardware manufacturers, it is much more reliable to explore AI applications in combination with their own basic business than those companion robots.
Is it good to have a cool back against AI?
Compared with home appliance companies that are obsessed with embodied intelligence, startups are much more ugly.
The reason why the state of these companies is summarized by "eating appearance" rather than "brain hole" is because the author is very sureThese entrepreneurs are fully aware of how nonsensical their products are.
For example, the team of the oven launched by a start-up called Seergrills said that through the precise control of heat radiation by AI, the oven can complete the grilling of a steak in two minutes at a temperature of 1,000-1,600 degrees Fahrenheit (about 538-871 degrees Celsius).
***cnet
It's hard to imagine how proteins can avoid full caramelization at this temperature, and while AI can empower all walks of life, it shouldn't reverse the laws of physics.
In addition to this completely "out of tune" AI application, there are also some ideas that seem to be pretty good.
For example, the handheld AI hardware product "Rabbit R1" released by the startup Rabbit.
This product comes with a piece of 288-inch touchscreen, a rotating camera for shooting, a scroll wheel button for interaction, and a microphone.
According to the company, the Rabbit R1 can provide users with services such as navigation and online car-hailing through voice input. The Rabbit team also highlighted that the product is based on a "large action model (LAM)", which means that after the user gives instructions in natural language, Rabbit R1 can understand the user's intent and complete the corresponding task.
***tom’s guide
In terms of the operating logic of the system alone, this device solves the problem of many mobile phone manufacturers when designing UI OS - reducing the user operation layer, and it can even be said to achieve a "zero operation level", because the user does not need to call the APP at all, he can give the corresponding instructions, and this device can also be used independently through a SIM card or WiFi link network.
However, the Rabbit R1 is definitely not filling a gap in the market, as it simply does not occupy any niche in the end market.
One of the most immediate questions is, can this handheld AI hardware replace mobile phones? The answer is obvious: not at all. So why do consumers carry another smart terminal in addition to their mobile phones?
Secondly, it can be seen from the design concept of the rabbit R1 that this device relies heavily on voice interaction, and the above block 2The 88-inch touchscreen probably can't support other ways of interaction, but the problem is that consumers rely on voice interaction very low, and it can neither replace touch interaction nor meet the privacy needs of users in public places.
This is actually very easy to explain,Why in today's mobile phone manufacturers"Roll up and die",Whether it's Siri,Or the voice assistant of domestic mobile phones,It's consistent and difficult to use,Because users really don't have any demand for this function。
We can further imagine that even after the access to the large language model, the voice assistant will be qualitatively improved in terms of functional implementation, but does this require the participation of other intelligent terminals? Mobile phone manufacturers can deploy large models locally based on the powerful computing power of mobile phones.
Taking a feature that is already integrated on a smartphone and selling it separately is no less magical than launching a PHS in 2024.
Annihilated? Having said all this, should all hardware products with AI as a selling point be killed with a rod? Of course not,As mentioned above, the TV that uses AI chips to improve picture quality,As well as the AI PC laid out by various OEMs,It's a good attempt。
In fact, looking at the electronic products that used AI as a selling point in the past, there are traces of their success path.
At the Tiger Sniff Generative AI Salon that just ended on January 27, Yibo, the founder of Yichuang Technology, mentioned that the real value of AI lies in changing the relationship between supply and demand, that is, providing scarce functions or services, rather than simply improving production efficiency. In the application of the previous era, the simple integration of AI functions was actually limited by the cognitive level.
So what kind of products have the attributes to provide scarce services? A typical case is the smart speaker represented by the Amazon echo.
The significance of echo at the product level is to simplify tedious interactions such as querying the weather in the home scene into voice commands at one time. And in the home scenario, the user can say voice commands without causing embarrassment, which is very different from today's AI portable devices.
Another interesting case is the Smart Fitness Magic Mirror. The level of financing for practitioners of this product in the past three years has been quite exaggerated, and in 2021 alone, Fiture raised $300 million. The scarcity of its AI level lies in the fact that everyone's fitness movements will have a variety of shapes, and real coaches will guide them according to the specific situation, which traditional fitness courses cannot solve this pain point. This is the core value point of the fitness mirror to develop the C-end market.
In fact, in the past, there was almost no case of a 3C product that rose from the original penetration frustration and rose only by integrating AI functions.
Another problem is that AI must not only provide more product value on the C-side, but also have enough resources to leverage on the B-side, otherwise it will be a bit "alone".
For example, the emergence of smart speakers has actually broken the situation that smart homes have appeared in the form of single products in the past, and have become the entrance for users to cut into IoT through functions such as voice recognition and device central control, and then leveraged a large number of related industries and formed a linkage development.
Although according to IDC statistics, since June 2021, the global smart speaker industry has entered a stage of continuous contraction. But in those years when business was booming, the Echo's explosive sales had completed its historic mission as a commercial product.
At present, those who have not yet found application scenarios, and strive to artificially create demand for usersAIInter, do not have the basic disk of smart speakers back then, so the answer may be obvious to whether these products can be favored by the market.
People who are changing and want to change the world are all on Tiger Sniff app