An Attempt to Review the Rabbit R1

I was pretty excited about mine Rabbit R1 Test device and planned to try it out for a whole week. My plans included using it outside around town, using it around the house as an all-day companion, and playing around with it with my friends. Unfortunately, I quickly ran out of things to do with it. It currently offers little that is worth testing.

Rabbit’s promise

In his Keynote that came around CES 2024, CEO Jesse Lyu promised the world with this device. He insisted that we don’t think of it as a phone replacement, but rather as a gadget that complements the way we use our phones. According to its makers, the R1’s sole purpose is to save time and maximize convenience by reducing the number of taps on our phones.

The idea is that the R1 makes a one-time connection to the services you regularly use and then handles them for you. This eliminates the need to use a different, dedicated app on your phone for each service. The R1 currently supports Spotify, Uber, DoorDash and Midjourney. Rabbit aims for a future where we no longer have to navigate through five different apps on our phone to perform five different tasks. Instead, the R1 is intended to act as a kind of automated personal agent. One that is capable of completing (some of) our everyday tasks – calling a ride, ordering food, playing music and more (?) – with simple voice commands.

Lyu also promoted the gadget as a quick, standalone solution to getting things done, unlike our phones, where we would often get distracted by a notification or the lure of social media and end up wasting half an hour on a three-minute task. An idea that is promising.

You can read the details on how it is supposed to work Herebut the idea is to go a step beyond the LLM (Large Language Model) that some apps have provided in recent months and introduce a LAM (Great action model). This means the R1 can do more than understand and respond to you; It can also perform actions for you, which should theoretically make it smarter than the voice assistants on our phones.

Another feature Lyu boasted about is the R1’s ability to understand natural language and filler words, which he said gives it an edge over other AI chatbots. It also guaranteed and demonstrated context awareness, which meant users could ask follow-up questions about a topic. On paper it all sounded interesting.

The design and hardware are attractive for the price

What I like most about the R1 is definitely its looks. Designed by Teenage Engineering, the R1 has a retro, old school look that will evoke nostalgia for all 90s kids. However, the general reaction from online users seems to be split on their opinion of the bright orange. Some praise it, others say it hurts their eyes. I agree that it’s a lot brighter than it looked on the screen, but I’m not complaining. I praised the recently introduced bright yellow Nothing sprouts because of their eye-catching color, and I feel the same way about the R1. I get bored when technology is black, white or silver.

When it comes to build quality, the R1 delivers exactly what you pay for. Although everything is made of plastic, the case is well made and feels pretty sturdy. It’s neither light enough to look thin nor too heavy to feel like a burden in your pocket. Plus, it’s a perfect size and fits easily into my small jeans pocket. But remember, you won’t lose your phone while carrying this thing around.

The 3-inch screen makes some users wish it was fully touch-enabled, but I like the R1’s current controls. In an age where everything from laptops to watches have a touchscreen, the R1 feels like a breath of fresh air and takes me back to the era of buttons and scroll wheels. However, you can tap the screen while typing on the R1’s keyboard. Otherwise the display does not accept touch input.

I also appreciate how minimalistic and easy to navigate the controls are. Only these two controls – the push-to-talk button (PTT button) and the scroll wheel – are responsible for everything on the device. For example, the PTT button is pressed once to lock/unlock the R1, long pressed to talk to it, and pressed twice to turn the Rabbit Eye on/off.

The R1 responds with sound effects and haptic feedback when it activates its vision function or types on its keyboard. Both editions helped make the user interface smooth and seamless. In terms of user interface, the R1 is almost there, but falls short in one crucial area. When you talk to the device, an eager rabbit appears on the display, bouncing up and down with its ears perked up to show that it is listening to you. As soon as you have finished speaking, the ears will be lowered again.

When you use the Vision feature, the R1 will also let you know that it is working on your request to keep you updated on the status of your request. But if you talk to it outside of the vision function, it doesn’t show any status updates, which is strange and I hope they get added soon. When I asked the chatbot a random question, I would often repeat myself and then realize that it was actually processing my request and I had interrupted it, meaning I had to start over.

Everything else sucks

When I reviewed the R1, I felt like I was using an empty orange box with nothing inside. Aside from its aesthetic qualities, there’s little to appreciate about this device. All features and functions are promised for later this year.

I wrote a separate story that I delve into in depth Everything I hate about the R1, but here’s a summary: It’s an unfinished product. Everything it currently offers is extremely simple and nothing we can’t do on our phones. Sure, our phones aren’t capable (at the moment) of what the R1 could possibly do in a year, so it probably should have been released in a year.

Even with the few services it currently offers, it struggles. It often misunderstood my wishes, acknowledged them but did not carry them out, or ignored them. Uber “worked” on the second try, but got both my pickup and drop-off locations completely wrong the first time, and Uber is the last app I want to risk my order with.

Spotify was a complete mess. Song and artist names were misunderstood and my personal playlists were not recognized even though I was logged into my account.

Doordash worked, but the R1 is the last device I would use for this app. My food ordering process is rarely as simple as a single voice command. I have tons of options to sift through and I prefer my phone’s 6-inch screen for that.

It took Midjourney a day of troubleshooting with the folks at Rabbit to get it working, and when it worked, it was… Midjourney. I don’t see the point in creating images on this tiny handheld device when I can do it on my laptop in much better quality. The inclusion of Midjourney in this gadget doesn’t make much sense to me. But if you really want to use your voice to generate random images on the go, you might be exactly the customer Rabbit is looking for.

The vision function wasn’t anything special either. It’s far less impressive than Google Lens when it comes to describing an object you point the camera at – and Google Lens is six years old. When it didn’t answer my query completely wrong, it described it so vaguely that it wasn’t very useful.

Additionally, RabbitOS experienced inexplicable disconnections, wasting a few seconds of my life each time I waited patiently for the R1 to fix its problems and execute my command.

In addition, it is characterized by a surprisingly short battery life. It’s almost comical that a product that is marketed as an “all-day companion” doesn’t last longer than five to six hours.

Check back in a year before you buy this thing

The R1 still needs a lot more work to become the next big thing, and Rabbit has a tight schedule. It already has has disappointed many users, and the company has to work quickly to deliver what it paid for. Rabbit also needs to work on what services it offers on the R1 and find a design for the app menu that makes sense for a 3-inch handset.

I’m being generous with the two stars here. Although it failed miserably at almost everything it set out to do, I believe the concept does have potential. The design and construction are already clarified and the price is so low that some people will be willing to try the experiment. If Rabbit is fast enough to figure out the functionality it originally promised, it could be an exciting device.