The Rabbit R1 offers an easy-to-use push-to-talk button and onscreen keyboard to navigate its generative AI, along with a SIM card slot to support 4G LTE connectivity.
Lyu promises to complete tasks such as calling Uber, reserving dinner through OpenTable or playing music on Spotify with just your voice, all with just an audio prompt from you. Lyu claims it learns these tasks by watching how humans use apps.
What is Rabbit?
Rabbit is a compact AI assistant that integrates with apps and is controlled through voice-command technology. Featuring an orange case and analog scroll wheel, as well as a 2.3GHz MediaTek processor and 128GB storage, plus push-to-talk button and camera; its company claims it can answer philosophical queries, check stock prices and play music via Spotify while translating text and taking dictations notes via voice control – in its CES keynote keynote, Rabbit CEO Jesse Lyu even used Rabbit to solve an office problem by scanning paper spreadsheet and digitizing it digitally using Rabbit!
The Rabbit R1 stands out from other smart speakers or AI assistants by not needing to connect it with your phone or tablet in order to use it, and is capable of doing things no other AI can match. It utilizes an innovative AI called Large Action Model which navigates apps and performs tasks much better than current AI assistants that often remain limited in what information they can give or only provide basic responses.
Lyu demonstrated Rabbit at CES as performing various tasks such as ordering from Doordash and Spotify, answering philosophical queries, looking up movie tickets and checking stock prices – but many reviewers criticized its usefulness given the price point; tech influencer Marques Brownlee described it as only “barely reviewable”.
In practice, Rabbit isn’t much more powerful than a smartphone or tablet; you can only access four services — Uber, Spotify, Midjourney and Doordash — through Rabbit’s website, which feels unsecure. Further, its training capabilities remain unknown so most of its functionality requires having your phone nearby to use.
Rabbit’s team has always maintained that this R1 model represents only the beginning of its journey and they recognize it needs further development to be successful. But for now, its release feels like bait-and-switch tactics aimed at coaxing thousands of people to pay $199 so that beta testing may begin.
The Rabbit R1 is a smart assistant
Rabbit claims its R1 can perform tasks such as playing music from Spotify, calling an Uber ride and planning a vacation (including flights and hotels), translating conversations, recording voice memos and making phone calls. With its bright orange case and simple design reminiscent of Pokedex or Google Bard devices, its powerful hardware system includes a 2.3 GHz MediaTek processor, 4GB RAM memory and 128GB storage capacity.
Rabbit R1’s Large Action Model, or LAM, serves as its brain. Trained on screenshots and videos from apps like Uber and Spotify as well as web browsers to navigate their interfaces and perform tasks just like humans would, it knows what an icon for settings looks like and when an order has been confirmed; and knows where search menus are located on websites. According to Lyu, its LAM can even teach itself specific tasks–for instance removing watermarks using Adobe Photoshop.
For this, the R1 connects to apps through Rabbit Hole portal that lets users log in and authorize services. However, downloads of apps won’t work directly through it and some services won’t support using it (this might work fine for some users; but not everyone will find this acceptable).
Additionally, the R1 lacks an “always listening” mode; only using its microphone when you press its push-to-talk button and activating its rotating camera when pointed at something it recognizes as human faces (good for privacy and security).
But the device remains under development, and its effectiveness remains uncertain. One major drawback to using it is requiring Internet access and a data plan from your wireless carrier if you want to take advantage of this device. Furthermore, responding vocally could become annoying in noisy environments.
The Rabbit R1 is a smart speaker
The Rabbit R1 is a colorful square gadget with an analog scroll wheel, two microphones, a 2.88-inch touchscreen and what CEO Jesse Lyu refers to as its “360 degree rotational eye.” These features come together for an experience unlike anything you’ve seen before!
To use it, press and hold a push-to-talk button or give voice commands and see an animated rabbit head bobble in response. Once your task has been given to it, its LAM powered by Perplexity AI gets into action to carry out your request – going one step beyond chatbots that use Large Language Models that simply respond with text based responses; LAMs take this further by acting like software agents capable of performing tasks on apps and screens without complex integrations such as APIs.
Rabbithole allows users to connect the R1 with services like Uber, DoorDash, Midjourney and Spotify, making the connection simple and secure. I found signing in a little unnerving at first despite their claims of protecting credentials securely.
This device can perform some useful tasks with its smart features, including automatically ordering food from DoorDash and providing you with a list of restaurant options. It also plays music through Spotify and answers simple questions like the distance of one mile or who sampled Isley Brothers’ “That Lady.” Furthermore, its built-in camera recognizes objects such as headphones, cars and dogs and can take notes through its built-in speaker system.
At the launch keynote, Lyu promised that additional functionality would be added to R1 in due time. He mentioned work being done on travel queries as well as plans for third-party reservations and ticketing support, smart home functionality, shopping services and teach mode, which allows you to demonstrate to Rabbit how to complete a task within an app or service. Lyu plans on eventually including video messaging as well as third-party payments as additional services that R1 may support in its future evolution.
The Rabbit R1 is a smart device
The Rabbit R1 is a compact AI assistant that integrates seamlessly with applications and can be controlled with voice-commands. Its design combines elements from smartphones and smart speakers, offering minimal interface with screen for navigation of apps and its artificial intelligence learning your preferences to provide customized recommendations and improved functionality.
This device uses a Large Action Model (LAM) operating system to quickly navigate apps and complete tasks more quickly than smartphone. It supports several services like Uber, Spotify, Midjourney and DoorDash; you can link these through an online portal. In addition, Rabbit R1 can learn to recognize certain website screens so it can perform clicks or key presses for you.
In practice, however, the R1 can be somewhat disheartening. It sometimes misreads your questions, for instance if you ask whether the Celtics are playing tonight it might say no even though the game has already started! Unfortunately this device serves a similar function to Siri or Alexa so hopefully future updates will improve its performance and reliability.
The R1 is an eye-catching orange cube designed by Swedish firm Teenage Engineering, responsible for designing such consumer electronics as Rigado robot and other consumer-friendly gadgets. Its key feature is an interactive scroll wheel which lets you access its menu of services – much like using a walkie-talkie!
R1 serves its primary function by responding to voice queries, but can also perform other limited tasks. You can set alarms, check calendar events and listen to music as well as connect to Wi-Fi networks and display notifications on its screen. Furthermore, its USB-C port makes charging convenient.
The R1 holds great promise, yet has yet to demonstrate its worth as a smartphone replacement. Unfortunately, certain features — such as being unable to set an alarm or view your calendar — remain absent on its platform; in 2024 they should be standard on every connected device.
