Los Angeles-based AI startup rabbit inc. today unveiled their new mobile device called r1, which aims to replace smartphones with a more natural, voice-driven interface powered by AI.
In an online launch event at CES 2024, rabbit’s founder and CEO Jesse Lyu introduced r1 as the first realization of the company’s vision for “the simplest computer” - a device so intuitive that no learning is required.
“The problem with [smartphones] is not the hardware form factor, it’s what’s inside - the app-based operating system,” said Lyu. “Each time you want to do something, you fumble through multiple pages and folders to find the app, and there are always endless buttons to click.”
To solve this, rabbit has developed a completely new AI Foundation Model called the Large Action Model (LAM), along with a custom OS called rabbit OS. While Large Language Models like ChatGPT understand language, LAM goes further by actually taking actions on a user’s behalf.
Powered by LAM, r1 replaces app icons with natural voice commands. Users can say things like “Get me a ride home” and r1 will book a rideshare without any typing or taps. It can also handle complex multi-step actions like planning an entire vacation.
“LAM can learn any interface from any software regardless of platform,” said Lyu. “In short - the Large Language Model understands what you say, but a Large Action Model gets things done.”
Physically, r1 sports a unique design with a rotating 360 camera for computer vision, a push-to-talk button, and an analog scroll wheel. It supports 4G LTE networks for global connectivity.
Industrial design comes via a collaboration with Swedish firm Teenage Engineering, known for its iconic synthesizers.
On the software side, r1 runs the Linux-based rabbit OS. A portal called rabbit hole lets users select and log into third party apps and services they want to activate, like Spotify, Uber, or food delivery.
Importantly, Lyu emphasized rabbit’s commitment to user privacy: “We don’t create fake or spam users, hack infrastructures, or store third party credentials.”
For authentication, r1 redirects users to actual login pages for each outside service. No user data gets tracked.
Rabbit also introduced an advanced “teach mode” where users can demonstrate workflows for r1 to learn new skills over time.
So when will we get our own r1 companions? Devices can be pre-ordered now on rabbit’s website for $199 - a fraction of high-end smartphones - shipping early 2024.
The launch caps off a monumental 12 months for rabbit, which raised an additional $10 million in December to bring its total funding to $30 million, led by Khosla Ventures.
Rabbit is betting that after over 15 years, the app-based paradigm is ready for disruption by this new generation of AI-powered devices - with r1 just being the first rollout.
“In a decade, we could have tens of billions more agents than people on the planet running around the net doing things on our behalf,” said Vinod Khosla of Khosla Ventures about LAM's potential.
If r1 delivers on its promises, our smartphones may soon be relegated as entertainment-only devices as AI companions like this handle everyday tasks for us. Rabbit believes the simplicity of voice is the interface of the future.
The race is on to see if startups like rabbit can define the next era of computing - delivering a more helpful, responsive agent in every consumer’s pocket.