San-Fran's Humane launched a wearable AI gadget called the ai Pin. . Now tech startup Rabbit has partnered with Teenage Engineering for a bright orange personal assistant for your pocket called the R1.
Think of the R1 as a pocket-sized AI assistant designed to take over those repetitive digital tasks we all find tedious. Unlike phones that are app-centric, the R1 doesn't display apps nor does it connect to app APIs. Instead, it uses AI to execute tasks. You interact with the R1 through voice commands, initiated by a push-to-talk button. This ideally allows for a more focused and efficient execution of tasks without the distraction of multiple apps.
It runs Rabbit OS, which uses a combination of large language models (powered by OpenAI’s ChatGPT) to understand your intent and large action models developed by Rabbit. These AI technologies will carry out your requests and can even learn by demonstration - they remember how you perform a task in an app and then replicate that task when asked. Rabbit said it has trained several sequences for the most popular apps, and Rabbit R1's abilities will grow over time.
Where existing voice assistants are limited to simple information requests or basic control of smart home gadgetry, the R1 extends such things to "handle most of one's digital errands – from simple tasks like searching for up-to-date information to complex tasks such as thoroughly researching and booking options for upcoming travel, or filling a virtual grocery store cart and completing transactions at check-out.
Rather than fumbling through apps on your phone, you can simply tell the R1 what you need, and it takes care of it. The Rabbit OS can manage various tasks -- like controlling music, ordering a car, buying groceries, and sending messages through a single interface. The interface displays category-based cards for different services.
R1 runs Rabbit OS. It's powered by a blend of large language models, including elements from OpenAI's ChatGPT, and Rabbit's proprietary large action models (LAMs). Instead of relying on APIs and developer support, Rabbit has trained its LAMs by human interactions with existing apps, teaching it to recognize and use apps.
On the right edge is a push-to-talk button that you press and hold to give the R1 voice commands. There’s a 4G LTE SIM card slot for constant connectivity, too, so it doesn’t need to pair with your phone or another device. You can connect the R1 to a Wi-Fi network, as well.
One of the R1's most interesting features is its "teach mode." You can demonstrate a specific task to the device, such as a sequence of actions in an app. The R1's AI then learns and memorizes these actions, enabling it to replicate the task autonomously in the future. This feature basically personalizes it to your specific needs.
Described as about the same size as a stack of Post-It notes and tipping the scales at 115 g (0.25 oz), the standalone R1 is home to a 2.88-inch touchscreen display rocking category-based cards representing different LAM-assisted activities, and a scroll wheel. Much of the interaction will be by voice command via the push-to-talk button (the device does not have an "always listening mode") and there's a rotating AI-enhanced camera for making video calls.
It comes with both cellular (via an optional SIM card) and Wi-Fi connectivity, has 2.3-GHz MediaTek Helio processing brains supported by 4 GB of RAM and 128 GB of storage, and the battery is reckoned good for all day usage.
The Rabbit R1 is up for pre-order now, is priced at US$199 and is expected to ship from early March. Global sales will follow later in the year.
Also Read: Steelbird amazing helmet launched