Another word about Rabbit R1

by Dmitry Kirsanov 7. March 2024 14:07

It’s only weeks until a revolutionary gadget to be released, so for the sake of history, let’s join the pieces of information we have regarding this prototype device. I’m saying prototype, because it’s the first of a kind, and even if it will fail, the idea behind it is what makes it a piper of new industrial revolution.

Disclaimer: I’ll quote Rabbit team below, so when I say “we” it means “they”. All of this information is first-hand and is not my opinion. It’s not necessarily true, but what the Rabbit declares as truth. It’s here to compare with reality later, when first reviews will be published.

The most commonly used example is DoorDash, U.S. food delivery service. That’s because Rabbit representatives are quite hungry until their devices become popular, take over the market and allow them to dictate the rules.

The Rabbit device, priced at only two hundred dollars, is a surprisingly affordable piece of modern technology. What's even more surprising is that there are no subscription fees attached to it. Rabbit has confirmed that they have no plans to introduce any subscription fees, although this could potentially change in the future. The reason for this affordability is the efficiency of the large action model and the Rabbit OS, which are less expensive to run than you might think. Once you purchase the device, it's yours to use without any additional charges.

There seems to be some confusion about the difference between a large language model and a large action model. Many people assume that it's just a marketing term or a large language model with a few extra features. However, this is not the case. A large language model, based on the Transformers architecture, is excellent at predicting the next token in a series of tokens, or the next word in a sentence. But this wouldn't have worked well with the Rabbit application.

What Rabbit does is convert your natural language into the action you want to accomplish, and then performs that action on a hidden operating system. Imagine using your phone to order food from a restaurant via DoorDash. Rabbit does all of that for you, figuring out how to navigate through the interface. It doesn't connect to an API, which makes it more efficient and less costly to run, hence the lack of a subscription fee.

Jesse Liu, in a podcast, explained the difference between a large language model and a large action model. He clarified that Rabbit is not creating a large language model, which is based on Transformer and requires a significant amount of GPU on the cloud to train and get things right. Instead, Rabbit works with the best language models and small language models, and potentially open-source language models in the future. They continuously monitor the performance of these models and can switch between them as needed. This is how Rabbit OS works.

Rabbit focuses on the large action model, which is neurosymbolic. While the Transformer-based language models were designed to understand language better, they perform poorly at completing tasks. Rabbit doesn't work with APIs because not everyone will provide an API, and even if they do, they often don't replicate the full feature of the application. Instead, Rabbit aims to create a universal solution, an AI that can handle any application, whether it's an Android app, iOS app, or Windows app. I think it’s important to note, that “Windows” in this case is the same as Linux was for early .NET. In 2001, Microsoft did think that .NET will be cross-platform, but didn’t do a thing about it. Same is here, Rabbit isn’t “Windows-first”, but of course the LAM will be compatible with Windows. Same as LLM is not just Python and NVidia. Did you know that it’s not?

To achieve this, Rabbit started working with data labelling companies about two and a half years ago. They assembled test groups to interact with different software like Uber and Spotify, and recorded these interactions. These recordings were then fed into the large action model, which reads the clips frame by frame.

Many people have questioned why Rabbit couldn't have just been an app on a phone instead of a standalone device. The answer is that Rabbit is trying to break the mold and show us the future of computing, which they didn't believe could be achieved with just an app. When you use your phone, you navigate through apps and find what you need to do using your finger. But with Rabbit, you speak in natural language and an AI agent performs the actions on your behalf. This wouldn't have been possible if Rabbit was just an app.

As for the concern of having to carry around two devices, it's worth noting that many people already carry around multiple devices.
In today's world, we are surrounded by various devices - phones, laptops, earbuds, and now AR and VR devices. As an early adopter, I find it exciting to carry around another device, especially one that is sleek, small, and lightweight. This device, which I believe represents the future of computing, is something I don't mind carrying around.

Many people wonder if Siri and Google Assistant will eventually incorporate the functionalities of this new device. While it's likely they will add some of these features, they've had over 15 years to do so and haven't quite hit the mark. Things are bound to change rapidly in the future, and I'm excited to see some competition in the market.

Apple and Google are trying to avoid the pitfalls of the innovator's dilemma. They have existing products with established frameworks and ways of thinking, making it difficult to break out of that mindset. Even if Siri and Google Assistant evolve, they will still have to work within the constructs of what they've previously built.

Rabbit, on the other hand, wanted to innovate and do something completely new. Jesse Liu, the creator of Rabbit, explains why he decided to build his own hardware device instead of just creating an app. He believes that the Rabbit is faster and more intuitive than other devices. For instance, when he needs to define a word or look up some information, he finds it quicker to use the Rabbit than to open a new tab and start typing on his phone or laptop.

The Rabbit also has a music feature, which Liu describes as a second iteration of the classic iPod with even simpler controls. He uses the Rabbit for music and searches most frequently, with searches making up about 70% of his usage.

Privacy and security are also major concerns for many users. Rabbit takes these concerns seriously and has a dedicated page on their website about privacy. Like, every other website and company. The device has a physical way to block its camera and does not have an always-listening mode. It only hears you when you hold down the physical push-to-talk button. The Rabbit OS also features a gyroscope-enabled privacy mode where all modes of input and output are disabled when the device is facing down.

As for the information you share with the Rabbit, the company ensures that it stays with the device. They will not use personally identifiable information for any purpose other than serving you. You can choose to have Rabbit OS remember or forget things through their Rabbit Hole web portal. They are committed to never selling or sharing your data with any third party without your explicit permission. This commitment to privacy is crucial in today's digital age.
Rabbit OS is a secure and innovative system that allows you to interact with your own Large Action Model (LAM) in a dedicated and isolated environment on your cloud. When you use Rabbit OS, you are assigned this environment, which is solely for you and your agent.

When you need tasks performed, Rabbit will use your own accounts, which you have “securely granted them control” over through the Rabbit Hole web portal. This could be for services like DoorDash, Spotify, or any other service. You securely log into the web portal using SSL, providing authentication permission just like you would with any other third party.

We prioritize your security and privacy. We do not store your passwords for these devices. During the execution of any tasks, especially those involving sensitive actions such as payments, Rabbits will ask for your permission and clarification. This ensures that you won't accidentally order $500 worth of Snickers bars from DoorDash.

Rabbits are a safe and respectful representation of a legitimate user. We do not create fake accounts, spam accounts, or engage in any abnormal access patterns or traffic to the target app. The AI agents we use look and behave like humans. We do not attempt to reverse engineer any service or aim to replace existing API equivalents. We do not solve captures and we do not seek to significantly alter the user base of the apps.

The agent performing an action on your behalf should be no different from you performing it yourself. The only difference is that it saves you a ton of time by doing it for you.

Rabbit OS is a startup, constantly evolving and improving. We strive for transparency and innovation.

Your passwords are not stored on the device. All security measures are done securely in the cloud. The hardware device simply acts as an interaction method between you and the agent.

Jesse Liu, a representative of Rabbit OS, explains how authentication works. When you tell Rabbit to order sushi for a family of five, for example, it knows how to use the Uber Eats or DoorDash app. This request is then sent to your service somewhere in the cloud. You have already authenticated your Uber Eats and DoorDash account in a web interface.

The device doesn't have any software pre-baked in or anything pre-installed. It's just an AI. You choose what kind of service you want to enable and how advanced you want this device to be.

We have a web portal that helps with all the authentication settings and feature management. You go to the website and choose whatever services you want to unlock. You then click the connect with Spotify button, for example. It will redirect you to Spotify login, and we don't save your credential.

We have a superhost computer that interacts virtually with a DoorDash app or website when you talk to your Rabbit R1 about ordering a hamburger from DoorDash. We then re-render Rabbit-themed UI on your device to get your result.

Conclusion and bets:

  • They are using hosts, and your device is not independent. That means – subscription is imminent. I’d estimate it to be the same as Amazon Prime or Disney+.
  • Big sharks will either buy them, or make their own “revolutionary devices”, depending on success of the Rabbit OS.
  • Rabbit is bound to become next Evil Corp in case of success.


Analytics | AI

blog comments powered by Disqus