The Rabbit R1 Is $199 and I Still Don't Want One

The Rabbit R1 Is $199 and I Still Don't Want One

Tech News ai-gadgets ces2024 hardware humane-pin rabbit-r1

CES is still winding down in Vegas as I type this, and the one product that's actually cut through the noise for me isn't a TV or a laptop, its a little orange square called the Rabbit R1. If you havent seen it yet: its a standalone gadget, about the size of a stack of Post-it notes, made in collaboration with Teenage Engineering (you can tell, it looks exactly like their radios and synths, same playful primary-color plastic). It's got a small screen, a scroll wheel, a push-to-talk button, and a rotating camera they call the Rabbit Eye. You talk to it, it does things for you. And it costs $199, one time, no subscription.

That last part is the whole story, honestly. Because the gadget everyone immediately compares this to is the Humane AI Pin, which got its big reveal back in November — a wearable clip-on thing that also wants to replace parts of your phone, except it's $699 up front and then $24 a month forever just to keep it working. Rabbit undercut that by a mile and got applause for it. Their first batch reportedly sold out within about a day of the announcement, which for a gadget nobody had touched yet, from a company most people had never heard of a week ago, is genuinely impressive.

Here's my problem though, and its not really about the price. It's about what these things are actually claiming to do. Rabbit calls the thing running under the hood a "Large Action Model" instead of a large language model, which is supposed to mean it doesn't just chat with you, it actually operates apps on your behalf, booking your Uber, ordering your DoorDash, playing your Spotify playlist, by having watched humans use those apps and learned the interface. No app store, no icons, you just ask for the thing and it goes and does the thing.

That is, if it works. And that's the part nobody at CES can actually verify from a demo booth. I've watched enough live tech demos over the years (I was writing this blog through the whole Google Glass thing, remember that) to know that a stage demo and a device sitting on your kitchen counter in April behave very differently. Demos are rehearsed. Real usage is you standing in a grocery store parking lot getting frustrated because it misheard "oat milk" as "boat film."

I also just have a basic skepticism about a whole new hardware category that exists to do things your phone already does, slightly differently, through a worse screen. Your phone already books the Uber. It already plays the Spotify playlist. The pitch here is that you shouldn't have to look at a screen full of icons to do it, which, fine, I get the appeal in theory. But I've tried voice assistants for a decade at this point and my honest experience is that talking to a machine and having it correctly interpret an ambiguous request is still the hard part, not the "no app icons" part. Siri and Alexa didn't fail because they had too many icons.

None of this means Rabbit is doomed, to be clear. $199 is a low enough bar that a lot of curious people are going to buy one just to see, myself included maybe, once units actually ship and reviewers get real hands-on time instead of a five-minute hallway demo. And credit where it's due: it's a genuinely fun-looking object, which is more than I can say for most CES hardware, which mostly looks like it was designed by a committee that hates joy. The scroll wheel alone is a nicer piece of industrial design than anything Samsung or LG put out this week, and those two companies have literal billions of dollars in design budgets.

But "fun-looking" and "actually useful enough to carry alongside my phone, not instead of it" are different bars, and I think a lot of the CES coverage this week is skipping right past that second one. I'll believe the Large Action Model stuff when I see someone use it to do something moderately complicated, unscripted, on camera, months from now — not at a booth with a rabbit-themed line snaking through the convention floor.

Anyway. Back to the usual programming around here soon, I've got a post half-drafted about migrating this blog's comment system off the ancient plugin it's been running since roughly 2013, which is its own small horror story.