Personal AI Is Here! Rabbit R1, Large Action Models and LLMs

If anything, last year was a bit of a shocker. We got not only a glimpse, but a real taste of the potential that AI can bring to the world.

Good, bad and ugly.

With it, just like any other technological leap, we saw silence (or ignorance), opportunists (or innovators), and rebellion (there are always torches and pitchforks).

AI, GenAI, LLMs, GLLMs, you name it, the world went crazy with the spread of super intelligent AI assistants. Whether you were a bigger fan of ChatGPT or Bard, the influx of new AI tools made technology that much more fun to play with.

This year looks like it’s about to bring us an even more personal, hyper- mega- supra-efficient AI. The best way to get this to the masses? Your smartphone. Pretty soon, you’ll find that smartphones are becoming genuinely intelligent, better at understanding your needs and acting on them. The key word being AND.

Action is going to be the highlight of the year, with proactive AI taking centre stage.

Trust us, it’s not a case of sci-fi fantasy, futurist ideology. It’s already here. With the Rabbit R1, fueled by their groundbreaking Large Action Model (LAM), Rabbit CEO, Jesse Lyu, sums it up perfectly: “Our mission is to create the simplest computer, something so intuitive… you don’t need to learn how to use it.” 

The Problem with Today’s Tech

Let’s face it, we’ve become very comfortable with the ritual use of our smartphones. 

They’ve hosted pretty much the same sort of operating system for the longest time, with a home screen, a selection of apps, a camera and some bloatware. Nothing truly innovative, except for bigger cameras, voice assistance and prettier displays, has been done for the longest time.

In his keynote, Lyu points out, “The smartphone was supposed to be intuitive… it no longer is.” It’s clear that while smartphones have evolved, their fundamental approach hasn’t. 

This is where Rabbit steps in with a solution. Something that can be seen as not just an improvement, but a potential revolution in the way that we interact with our devices.

Where Large Language Models Fall Short

You might be thinking that it would be a great new leap to just integrate ChatGPT or one of those powerful AI tools into the next wave of smartphones. But it still wouldn’t change much.

You could just get a standalone app that lets you use one of those LLMs, but it still requires you to navigate an entire array of applications to get your result.

Take your next holiday, for example. You can use an app with an LLM to determine what the most budget-friendly place would be, the best itinerary, historical points of interest, places to visit, etc. You still have to run that app and prompt the information, then you have to open your booking app and find flights, then you have to find and book accommodation. Nothing has really changed.

The process moved from the computer to the smartphone. That’s about it.

Indeed, LLMs are incredible. It’s machine understanding of our language. But they hit a wall when it comes to action. “LLMs… have made it much easier for machines to understand you,” says Lyu. However, understanding isn’t enough.

We can all agree for any real change to occur, we need action. And that’s where LAM enters the picture.

The LAM Difference

LAM is not just a fancier version of LLM. It’s a completely different beast. 

While LLMs excel at understanding language, LAM is designed to understand and execute actions. “The large language model understands what you say, but the large action model gets things done,” Lyu explains. 

LAM has been designed to observe human interaction with applications and replicate those actions. Which essentially surpasses the text-based limitations of LLMs. 

This is not just an incremental step; it’s a leap into a new realm of AI capabilities.

Rabbit R1: More Than a Gadget

The Rabbit R1 isn’t just another tech gadget; it’s the embodiment of a new AI philosophy –  one that promotes innovative use of AI to improve and enhance our day-to-day lives. 

The goal is to make technology a sidekick. A tool to augment and improve your livelihood.  

“Introducing R1, your pocket companion,” announces Lyu. 

The R1 is an intelligent device that uses LAM to seamlessly interact with various applications. It’s a device that listens, understands, and acts – a true companion in the digital age.

LAM in Daily Life

Back to that scenario I mentioned earlier. Imagine asking your device to book a ride, and it does so without the need for separate apps or unnecessary steps. Everything is integrated, interconnected and simplified.

“R1 can help me reach others just like a phone, but smarter,” Lyu states. The R1 can help you simplify complex tasks and turn a series of actions into one smooth interaction.

This new view on AI means having a personal assistant in your pocket that understands not just your words, but your intentions.

The Future of AI and Consumer Tech

The implications of the Rabbit R1 and LAM are vast. 

“With LAM fast evolving, my R1 will eventually help me to do things that can never be achieved on an app-based phone,” Lyu envisions. 

This isn’t just about making tasks easier; it’s about redefining our interaction with technology itself. The R1 and LAM are leading us into a future where technology adapts to us, not the other way around. It starts paving the way for a symbiotic relationship between humans and technology.

LAM vs. LLMs: What Sets Them Apart?

Is LAM just a marketing gimmick dressed in fancy tech lingo? According to the Rabbit website, LAM represents a paradigm shift in how AI interacts with applications. 

It’s not about understanding text; it’s about understanding actions and intentions.

“Language models are ill-equipped to comprehend applications with raw text,” Lyu points out. LAM, on the other hand, is designed to directly interact with the interface of any application, bypassing the limitations of text-based AI.

What Makes LAM Unique?

LAM’s uniqueness lies in its ability to model complex application structures, not just as simple token sequences but as interactive, functional entities. “We can define and model complex application structures… They are compatible with both a symbolic algorithm and a neural network,” says Lyu. 

This hybrid approach allows LAM to execute actions in a way that’s more aligned with human intention and understanding.

Learning Actions by Demonstration: The LAM Way

One of the coolest aspects of LAM is its approach to learning: imitation. 

It learns by observing human actions and replicating them. “LAM’s modelling approach is rooted in imitation, or learning by demonstration,” Lyu reveals. This means that LAM can adapt to changes and nuances in user interfaces, making it incredibly versatile and user-friendly.

Will It Alter the Way That We Interact with Technology?

What the Rabbit R1 offers is not just a new device but a glimpse into the future of personal computing. 

It takes on the status quo of app-based interactions, opening up a world where our devices understand us at a deeper level. 

It’s a step towards a more intuitive, efficient, and personalised tech experience.

This innovation is a new way of interacting with our digital world. 

The future is here, and it’s intuitive, efficient, and, frankly, pretty cool. 

With Rabbit R1 and LAM, we’re not just using technology; we’re interacting with it in a way that’s never been possible before. Welcome to a smarter, faster and fun future!

More in the Blog

Stay informed on all things AI...

< Get the latest AI news >

Join Our Webinar Cloud Migration with a twist

Aug 18, 2022 03:00 PM BST / 04:00 PM SAST