Table of Contents
You may have a virtual assistant on your phone or a smart speaker at home, but do you need them? A new startup called Rabbit thinks not. And the device is Rabbit R1. Rabbit believes these AI app devices are not the best way to use AI, and they have a different vision for the future. They want you to trade apps for conversation, and use a device that looks like a walkie-talkie to talk to an AI.
Meet the Rabbit R1 device
It is making its debut at CES 2024 in Las Vegas this week. The Rabbit R1 is a cute little square in a bright orange color, designed by the renowned Teenage Engineering. It has a small touchscreen, a scroll wheel, two microphones, a speaker, and a camera that can rotate 360 degrees. But the main feature is the “Push-to-Talk” button, which lets you communicate with the Rabbit OS, the AI behind the R1.
The power of the Large Action Model
The Rabbit OS is not like Siri or Alexa, which rely on apps or APIs to perform tasks. Instead, it uses the Large Action Model (LAM) technology, which allows it to take actions on any user interface. This means that it can learn to do almost anything you can do on a screen, by watching you do it once. For example, you can teach it how to generate an image using Midjourney via Discord, and then ask it to do it again later.
The Rabbit R1 can also handle common requests, such as booking an Uber, finding a recipe, or answering trivia questions. It responds with a stylized rabbit head on the screen, which bobs as it listens and talks. The camera can also identify objects or people in the real world, and suggest actions based on them. For instance, you can point it at a fridge and ask for a low-calorie recipe, and it will scan the contents and give you some options.
A more natural, engaging, and private AI
The Rabbit R1 is not just a gadget, but a new way to interact with AI. It aims to be more natural, engaging, and fun than the existing virtual assistants. It also promises to be more private, as the camera faces up into the body when not used, and the data is stored locally on the device.
However, there are still some questions about the R1 that need to be answered. How long does the battery last? How easy is it to train the Rabbit R1? How accurate and reliable is the LAM? And most importantly, do people want a walkie-talkie for AI?
The Rabbit R1 is priced at $199 and is available for preorder now. It is expected to ship in March or April. If you are curious about this device, you can check it out at CES 2024, where Rabbit is showcasing its potential and vision.
What is the Large Action Model?
The Large Action Model (LAM) is a new technology developed by Rabbit Inc. LAM is a type of foundation model, which means it can learn from data and perform tasks on any user interface, such as a screen or a camera. LAM can understand human intentions and behavior, and then mimic and perform them on various apps, without the need for complex integrations like APIs.
For example, LAM can generate an image using Midjourney via Discord, or suggest a low-calorie recipe based on the contents of a fridge. LAM is different from large language models like ChatGPT, which can only generate text in response to human input. LAM can also handle common requests, such as booking an Uber or answering trivia questions. LAM is the core technology behind the rabbit R1 device, which is a walkie-talkie for AI that was launched at CES 2024. The R1 device uses rabbit OS, which runs on LAM and has personal AI agents called rabbits that can carry out various tasks
