We’re getting very close to shipping the first batch of r1 devices from the factory. A ton of work has gone into the product since we first introduced r1 at CES in January. Our engineering and design teams have been working round the clock making updates to the product to add new features, improve latency and overhaul the UI. We can’t wait to see what you have to say when you try it for yourselves!
r1 is your ticket to experience the next generation of AI, where you have a personal assistant on hand to take care of actions for you. But it’s just the first step. We’ll keep working hard to consistently improve the experience over time. This is the beginning of a new journey, and we encourage you to grow together with us.
We know you still have some questions about what to expect, so we’d like to answer some of them here.
When will r1 start to ship, and when will I receive mine?
We’re proud to say that we’re on track for the first batch to start leaving the factory on March 31, just as we promised. It may still take a few weeks for the first batch of orders to be delivered because of customs and logistics. In case you’re wondering, there are a few steps in this process for batch 1 orders:
- The first batch of r1 devices leaves the factory and goes through customs. You’ll receive an email telling you that your order is being prepared for shipping.
- The devices get shipped to the U.S., where they go through another round of customs.
- After clearing customs, they arrive at our warehouse, where your order is processed.
- When your order is sent out to your delivery address, you’ll get an email with a tracking number.
- Finally, your order reaches your delivery address. You’ll get a final email confirming that it has been delivered.
Based on this process, we anticipate that the first r1 orders will reach customers in the U.S. starting April 24.
Please note that the steps above are for U.S. orders only, and the process may vary for future batches sent to other countries. Always remember to keep an eye on your order status for the most up-to-date information about when you will receive your order. For later batches, stay tuned to our official channels for updates about when those batches will be shipped.
What can I expect on day 1 after I receive my r1?
The following features will match what we demonstrated in the keynote on day 1:
Basic features:
- Conversation with LLM
- Up-to-date search with Perplexity
- AI vision
- Bi-directional translation
- Note taking with AI summary
LAM-powered features:
- Music
- Generative AI
- Rideshare
- Food
Regarding the travel feature, LAM is already capable of operating in the same way as we demonstrated in our keynote. We’re just working out some final regulatory details and aim to release it in a future OTA as soon as possible. We’re also working to bring more LAM capabilities with the most popular apps to be enabled on r1 via frequent OTAs. Check out the r1 page on our website to learn more about the status of current and upcoming LAM functionalities.
What is LAM, and how can I expect to see it develop over time?
LAM stands for “Large Action Model.” Simply put, LAM is a model that learns how to use any software it runs into, takes actions, and gets better over time. It learns by studying how people use online interfaces, and then it is able to operate those interfaces in the same way that a human would. Importantly, it also understands natural-language inputs, so you can speak to it like you would speak to a personal assistant. You ask it to do something, and it takes care of it for you.
Bringing a real-time agent, or “rabbit,” to a new kind of AI hardware is no simple task. While we have already been training LAM for a few years, we are taking a very careful approach in how we introduce it to the general public. For r1, we are first focusing on the most commonly used apps, building a custom UI for each category of service, and conducting thorough quality testing to make sure they all function well. Our aim is to create a delightful, efficient and intuitive user experience with privacy and data security at the core of everything we do.
The LAM functionalities listed above are just the first step in what will be a constantly evolving and quickly improving experience. As the saying goes, this is the worst this technology will ever be. Over time, LAM will offer many new features and greatly improved latency, meaning you’ll be able to accomplish much more in much less time. We will keep you updated on our roadmap as we continue to support more apps.
What is Teach Mode, and when will we see that?
Teach Mode is an experimental capability that lets you train your own “rabbits” to perform specific tasks on niche apps and workflows. This will give you the power to create your own personalized agents to handle tasks specific to your needs. And it will be simple to use, even if you don’t have any experience with coding or programming.
We are taking our time to make sure our platform is mature before we release this publicly later this year. We may invite insiders to try preliminary versions of Teach Mode, so if that interests you, stay tuned to our official channels for more information later this year.
How will updates work? Will I have to download OTAs?
r1 runs “headless” in the cloud, which means that most of the computing is offloaded to data centers. We intentionally designed it this way to move as fast as possible, and to bring updates and improvements to your r1 as they happen.
So what does that mean for OTAs? Unlike software updates on your smartphone, our version of OTAs will be more effortless. Many updates will happen in the cloud, so you won’t even see them happening. In essence, this means that you will wake up one morning and discover that your r1 is more capable than it was when you went to sleep the night before. Updates that require you to accept them will also be much lighter than a typical smartphone OTA (megabytes instead of gigabytes), and we’ll give you a clear prompt on your r1.
Our product team will be testing a lot of features and trying a lot of new things to make r1 a useful, reliable and exciting companion for the long-term. We’ll be improving things quickly and consistently, so your experience is constantly improving.
How does rabbithole work?
rabbithole is our portal to help you manage all of your connected services for rabbit OS and will be available from day 1. To connect to services that are supported on r1, all you need to do is sign into your account on the rabbithole, click the service you want to connect to, and sign in with your credentials.
Agents on our platform, which we call “rabbits,” can see the process of you logging into vendors’ apps. The act of entering your user name and password is encrypted and protected, and we do not store that information in our database. However, we do retain an authenticated state of your app on our cloud, securely stored, so that our agents can act on your behalf. You can remove this information whenever you would like to. We are also actively working with industry partners, including security auditors, to better secure your data.
What is rabbit’s approach to working with apps?
By connecting to your services through the rabbithole, you are granting access to allow our rabbits to operate them on your behalf. With this access, rabbits can see whatever you see on your apps. They may see private information, but we don’t store that information anywhere. A virtual environment is created on the cloud every time you ask r1 to do something for you, and that session is ephemeral (aka short and temporary) and discarded after your task is completed.
rabbits also respect consumer apps. Sometimes apps may refuse to provide service by blocking or giving you CAPTCHAs. To show our respect to those apps, we do not solve them, but rather accurately report them to our users.
And one note on performance: How well our rabbits perform on any given app is a direct result of the performance of the apps they are operating. Some apps are slow while others are faster. Some are more reliable while others are less so.
How is rabbit working with LLMs?
We collaborate with Large Language Model (LLM) vendors like Perplexity, Anthropic and OpenAI to use their services for understanding user intentions. When a user interacts with rabbit OS, we route their requests to the most suitable model capable of fulfilling their needs, whether it’s our in-house model or one provided by our partners. User data is stored on our servers; only processed, user-initiated conversations and utterances, which do not contain any personally identifiable, private, or sensitive information, are sent to our partners.
Note that LLMs, the underlying frameworks for understanding intentions, are still at their early stages. They hallucinate, make mistakes, and are not fast enough yet. We uphold a high bar of bringing the best experiences to our users. To this end, we are actively working with our partners to both push user experience and collectively define the capabilities and boundaries of LLMs.
What else is rabbit doing to help the development of this new industry?
rabbit is committed to giving back to the community by contributing to open research and engineering. At the moment, some of those areas include:
- Contributing to web automation and app automation frameworks.
- Participating in the discourse of agent interaction with consumer apps in the real world, by designing and releasing benchmarks, toolchains, and communicating with app developers.
- Contributing to on-device compute by releasing, in the future, optimized local models for perception and conversational AI.
In the same way that r1 is just our first step in rethinking the AI software and hardware experience, we’re also at the early stages of action-centric AI. We’re looking forward to doing even more to support the development of this industry in the future.
We really appreciate the support that we’ve received over the past few months. Even though we’re a very small startup with significantly fewer resources than the other giant tech companies out there, we’re still dedicated to delivering a solid experience with r1 on day 1. This is just the beginning.