The AI Pin by startup Humane was arguably the star of the show at MWC (Mobile World Congress).
As a fully AI-native product, AI Pin has met the long-held expectations for hardware innovation. With OpenAI's technology embedded, along with laser projection and gesture control, AI Pin appears to be an incredibly cool gadget. However, if judged by the standards of mass-produced consumer goods, AI Pin isn't quite there yet.
For instance, its laser projection is a manifestation of Humane's "screen-less" product philosophy, yet the display quality is akin to that of a Casio digital watch, far from matching up to the sophistication of mature OLED screens.
AI Pin Laser Projection | Image Source: Humane
Lacks localized applications and is heavily dependent on the internet. From the demo videos seen live, any AI response takes a few seconds of waiting time;
Priced at a steep $699 (approximately 5023 RMB), with an additional monthly subscription fee of $24 (approximately 172 RMB);
Setting these aside, the crucial question for AI Pin remains unanswered: Why do I need an AI Pin when I already have a smartphone?
The same question applies to rabbit r1, which made headlines at CES and sold 50,000 units in a few days.
Others have not taken such big steps. At MWC, German operator Deutsche Telekom, in collaboration with AI startup Brain.ai, showcased an AI concept phone, T-Phone.
They claim it to be a phone without apps—Could T-Phone be the prototype for future AI phones?
01, "App-less" Phone
In terms of specifications, T-Mobile can be considered a standard budget phone: with only 2GB of RAM and a Snapdragon 625 processor from 2016, 14nm, and dual rear cameras of 15 and 13 megapixels, respectively.
Its core concept is this: abandoning the app-centric design of phones, it uses AI to generate interfaces based on context prediction to help users accomplish tasks. Therefore, when you light up the screen, the first thing that appears is an AI assistant interface. There's also a physical AI button on the side of the phone for instant access to the AI assistant.
A typical scenario is shopping. When you ask, "What should I buy for my grandmother who is bedridden?" T-Phone will recommend relevant products, from "bedside tables" to "U-shaped pillows for neck support".
If interested in a product, clicking on it will generate detailed information including text, images, and videos. After selecting the desired item, clicking on the image will add it to the shopping cart of a third-party e-commerce platform.
Jerry Yue, founder and CEO of Brain.ai, mentioned that the system will attract about 7000 e-commerce sites upon official launch.
In this case, T-Phone acts as an e-commerce aggregator platform.
Another example demonstrated by Yue was purchasing airline tickets. He told the AI assistant to book two first-class tickets from Barcelona to Los Angeles for March 12. The phone paused for a moment, then generated a list of flights. Once Yue found a suitable flight, he could pay using his preferred mobile payment system without switching to any other app or service.
Using T-Phone to Purchase Airline Tickets | Image Source: CNET
Not just limited to shopping, users can instruct AI to perform various tasks, such as suggesting travel destinations, creating slideshows from photos, posting on social media, drafting documents, and more.
T-Phone's AI capabilities are based on a combination of cloud and edge computing. High-energy tasks are handled by cloud-based AI to reduce the phone's power consumption, while low-energy tasks can be processed directly by local AI to address delays caused by data transfer.
Brain.ai believes that today's apps, lacking the background of connections established by humans, cannot dynamically adjust or adapt to human needs in real-time. They also cannot learn from past connections to accommodate new requirements. Another drawback of apps is that they often aim to captivate users from the outset during development.
"We aim to use artificial intelligence to do the complete opposite – make humans the center of technology. You don't go to applications; applications come to you. We want to empower people," said Lesheng.
Deutsche Telekom and Brain.ai anticipate that in a decade, you won't even remember applications. All you'll need to do is converse with artificial intelligence, and it will handle all tasks for you.
LeSheng showcased the T-Phone at MWC | Image Source: androidpolice
Sounds promising, but currently unclear how to set up return and payment information; also, uncertain how the AI assistant will assist users when real customer intervention is needed for after-sales issues.
Brain.ai launched a similar AI assistant app called Natural in 2021, closely resembling the T-Phone concept. Despite a high rating of 4.7 stars, users pointed out issues like lack of after-sales support and higher flight ticket prices.
Moreover, what if users just want to chat on WeChat or check emails? In these scenarios, the addition of an AI assistant may not necessarily enhance efficiency.
02, AI as a Super Gateway
It's evident that T-Phone is still in its early stages, explaining its status as a "concept phone." Only a few usage scenarios have been officially demonstrated, and none are considered high-frequency. Trying to replace traditional smartphones with such a device appears challenging.
In reality, T-Phone is not a fully-fledged AI phone; swiping up on the main screen leads to the traditional Android interface. Essentially, T-Phone is an AI shell added onto a traditional Android phone.
This may be the smartest move in T-Phone's product philosophy, as it doesn't attempt to overturn existing smartphones, but rather positions the AI assistant as a secondary entry point.
Why is the entry point crucial? Because it's the first step for all functions to unfold. Every person passing through the entry point is a potential customer, possibly willing to pay a toll.
During the rise of personal computers, Microsoft seized the operating system as the entry point. Software built user scenarios, making hardware meaningful. At that time, all PC manufacturers were Microsoft users.
In the internet era, Microsoft gained control over the browser as the entry point, further solidifying the Windows operating system's monopoly.
Early internet users had a habit of writing down website addresses on paper, typing letter by letter to access them. Hence, portal websites like Yahoo emerged.
Subsequently, with the explosion of websites, searching efficiency on portal websites was low, leading to the birth of search engines like Google and the establishment of the commercial model for online advertising. Today, online advertising is the main source of income for Meta, Baidu, and ByteDance.
Similar to the T-Phone concept, devices like Rabbit R1 and AI Pin assume a scenario where everyone is eagerly ready to enter the AI era. This might be an illusion created by AI, as people might not be prepared for such a transition.
The birth of the iPhone wasn't spontaneous. Apple had experience manufacturing low-power devices like the iPod; flash memory needed to become feasible at an affordable price point; Samsung needed to produce high-quality processors; the 3G network had to be introduced; iTunes music store laid the foundation for the App Store.
Everything converged in 2007, sparking the explosion of the mobile era.
Technological iterations are often gradual, with a long iterative time between the emergence of new technology and its ultimate realization. New technology needs to develop sufficiently, and the contradictions with old technology must intensify before the courage to make a cut arises.
From this perspective, T-Phone isn't a phone that doesn't require apps. Instead, it heavily relies on the existing app ecosystem, making it more likely to succeed. People may not be ready to enter a world without apps, but many might be willing to reduce the number of apps on their phones.
