Rabbit, an artificial intelligence startup, has developed a device that, according to Jesse Lyu, its creator, is not designed to replace smartphones — although it could do so in some cases. Is called Rabbit R1, and it is a kind of assistant powered by a language model similar to ChatGPT with which users will be able to do much more than interact with a virtual assistant. Also, it is very reminiscent of Humane’s AI Pin.
The Rabbit R1 stands out in many aspects, including its price. The startup has confirmed that the device will cost $200; a considerable cost considering all its capabilities. On an aesthetic level, the Rabbit R1 is half the size of an iPhone, and its design is very minimalist. Specifically, it has a 2.88 inch screen. Also includes a rotating camera to take selfies or photos and videos, as well as a scroll wheel that also acts as a button to navigate the interface or use the device’s built-in wizard.
However, what is special about the Rabbit R1 is its software. It has Rabbit OS, a self-developed operating system based on what the company calls “a large action model.” This does not have applications as such, but rather works as a universal app controller. That is, with Rabbit OS, users will be able to access some of the services of their favorite applications, without needing to access them directly or log in. For example, they can order food delivery, make purchases, play music, make reservations, buy tickets, send messages, and more.
All of this can be done through the physical button, or through voice commands, as if we were talking to Siri or Alexa. The screen also shows relevant information about these actions as widgets or cards.
The Rabbit R1’s AI is capable of using the applications for you
Rabbit, on the other hand, does not use any API open to third-party developers so that they can adapt their applications to the device or operating system. Instead, they make use of their action model. The Rabbit R1’s AI is capable of knowing how to use the different applications on its own.
In fact, the aforementioned language model has been trained by real people who use platforms like Uber or Spotify, who taught the AI how they work.
The AI also functions as a dedicated language model. That is to say, the user can teach the Rabbit R1 how to do something in an applicationjust like Lyu explained “You’ll say, ‘Hey, first of all, go to a software called Photoshop.’ Open it. Select these photos. Make a loop on the watermark and click, click, click. This is how to remove the watermark.’” The model will learn the actions in a time period of 30 seconds.
The Rabbit R1 will go on sale soon. For the moment, those users who want to purchase it can do so from the company’s website. The device, we reiterate, has a price of 199 dollars.