Teenage Engineering co-designed Rabbit R1, an adorable AI-powered assistant. This dedicated virtual assistant gadget may fit in your pocket but it is not a phone. Yes, you undoubtedly have a virtual assistant on your phone. If you’re reading Engadget, I’m willing to bet you have at least one smart speaker floating about your house that you can use to do simple chores. However, a new firm named Rabbit believes that Siri and Alexa are less-than-ideal applications of AI.
Teenage Engineering co-designed Rabbit R1, an adorable AI-powered assistant
The business (which was present at CES 2024 in Las Vegas but was not an official exhibitor) envisions a society in which you trade apps for conversation and, rather than a distracting device putting icons in your face, you connect with an AI via what amounts to a walkie-talkie. We had the opportunity to see the Rabbit RI AI device for ourselves at the exhibition.
The Rabbit R1 is Rabbit’s first device, and it’s an undeniably gorgeous little square in an endearingly vibrant shade of orange. Even if you don’t believe a specific gadget for a virtual assistant is necessary, you can’t ignore the aesthetic appeal, which comes from Teenage Engineering’s design team. It has a modest 2.88-inch touchscreen, an analogue scroll wheel, two microphones, a speaker, and a “360-degree rotational eye,” which is simply a fancy name for a camera that can be rotated to face you or through the back of the handset.
The primary technique to engage with the Rabbit AI is to push and hold the “Push-to-Talk” button. This instructs Rabbit OS to start listening. A heavily stylized and disembodied rabbit head bobs slowly while you ask a question or assign a task, and then it immediately gets to work. Looking to book an Uber? Looking for a meal to use up the leftovers in your fridge? Wondering who sampled The Isley Brothers’ “That Lady”? (The answer is Beastie Boys, Basement Jaxx, and Kendrick Lamar, FTR.) The Rabbit R1 AI appears to be quite capable of performing such duties, at least in the controlled video demo.
Rabbit OS can handle those tasks via what it calls the Large Action Model (LAM). Jesse Lyu, the company’s founder and CEO, describes this as the company’s key invention. It is intended to perform operations on interfaces rather than through APIs or apps. In summary, it can be trained to perform practically any task that can be completed via a user interface. It’s similar to a fancy macro.
Lyu demonstrates the Rabbit AI’s skills by teaching it how to produce an image with Midjourney via Discord. Once Lyu walks and completes the process, Rabbit OS can repeat it when prompted.
The revolving camera faces up into the body by default, serving as a privacy shutter. When called, it just turns its sensor towards the target. It can perform standard tricks such as identifying persons or objects in the real world (within certain limits). However, the way it interacts with the AI will undoubtedly catch people’s curiosity. In the demo, Lyu points the R1 to a full refrigerator and asks it to recommend a recipe that is “low in calories” based on its ingredients.
Of all, there are still many unsolved questions about the Rabbit R1. How long is the battery life? The firm says it’s “all day,” but what exactly does that mean? Will the common user be able to train it easily? At least we know a few things. We know it costs $199 and is currently available for preorder, with a ship date of March or April.