[ad_1]
QUEENS—Rabbit chose the iconic TWA Hotel at New York City’s JFK Airport as the backdrop for its launch party. The company invited not only the media but more than 100 actual paying customers, all of whom schlepped out to Queens for the evening shindig. Rabbit used the swanky locale to distribute the first wave of customer orders, but, more importantly, showcase its CEO’s vision and what the Rabbit R1 AI-assisted device is capable of. Attendees spent the first hour of the evening mingling, sipping cocktails, and munching like rabbits on tiny hors d’oeuvres. Eventually, Rabbit CEO Jesse Lyu took the stage to raucous cheers to showcase the R1. He provided some background concerning the company’s approach to hardware (keep it simple, but cool-looking!) and how it’s handling user input and requests (lots of server-side number crunching). The goal? To stop using the apps on our phones to do things and start using actions. That’s why it relies on a Large Action Model (LAM) rather than a Large Language Model (LLM).
(Credit: Eric Zeman)
Most of the live demos worked, though a bit slowly due to the poor network service in the TWA’s underground event space. The R1 was quickly able to find the weather, scan a simple spreadsheet and rearrange the cells, hold a live conversation between two different languages via translation, and play songs from Spotify. When Lyu tried to place a food order through DoorDash, however, things came to a screeching halt and required a re-do. Lyu blamed DoorDash’s backend for the slow pace of progress. You can also hail an Uber (but not a Lyft) with the R1 as well as generate images through Midjourney. This took some time but still worked smoothly during the on-stage demo.
(Credit: Eric Zeman)
What’s critical to know here is that Rabbit is not using the SDKs or APIs or other developer tools from Spotify, DoorDash, or Lyft to accomplish these tasks. Instead, the LAM is at work behind the scenes running through the actual actions within the apps (or their associated websites) to perform the requests. This is what Lyu considers to be the game-changer. The issue? Training these actions for the millions of apps out there takes time, which is why the R1 works with only a handful of apps at launch.
(Credit: Eric Zeman)
As for the hardware, well, it’s cute. Teenage Engineering designed the Rabbit R1. The device’s square body is about half the size of an iPhone 15 Pro Max. It’s made out of plastic and includes a small display, a scroll wheel, an action button, a swiveling camera, a USB-C port for charging, a SIM tray for data service, and a small speaker. The bright red plastic design is endearing but somewhat blocky and slippery. The scroll wheel does not work at all like I imagined it would. Rather than move through on-screen menu items with a click, it rotates freely and often jumps past the item you intended to select. The button has a satisfying click, but learning the proper sequence of presses isn’t exactly intuitive. Though the screen is touch-sensitive, you can’t press items on the screen to select them. Instead, you have to scroll and click, like an old-school BlackBerry. Crucially, the device lacks any sort of “back” function to return to the previous screen our exit your current action. I think the biggest letdown is that you can’t speak your requests using a wake word as you can with Alexa, Google Assistant, or Siri; you have to press and hold the action button to perform actions. However, a simple shake will bring up the Settings menu, and you can activate an on-screen keyboard for pecking out action requests when speaking them isn’t an option.
(Credit: Eric Zeman)
Rabbit was smart when it came to the case. The device ships in a protective carrying case that resembles the case for an old cassette tape. When opened, the case doubles as a stand for the Rabbit R1. It’s a nice little ensemble to have sitting on your desk. The Rabbit R1 is available for order now. It costs $199. It supports 4G LTE and Wi-Fi for connectivity. You can snag a data-only SIM from the carrier of your choice for access when out and about. This pricing is in stark contrast to that of the $699 Humane AI Pin and its required $25 monthly service plan. The answer to Rabbit’s central question appears to have an answer, but not the one Rabbit wants to hear. Based on what we learned during the launch event and the few moments we’ve spent with the R1, the device cannot outperform our smartphones for simple AI-assisted tasks. The bigger question is, will we bother to put our phones down in the first place to see if the R1 is worth giving a shot.
Get Our Best Stories!
Sign up for What’s New Now to get our top stories delivered to your inbox every morning.
This newsletter may contain advertising, deals, or affiliate links. Subscribing to a newsletter indicates your consent to our Terms of Use and Privacy Policy. You may unsubscribe from the newsletters at any time.
[ad_2]
