The Rabbit R1, one of the first “AI gadgets”, is not unanimous among the first testers. Like the Humane Ai Pin, it lacks relevance to the smartphone, is inconvenient to use and contains many bugs. Worse, he would lie about how it works.
Does the future of generative artificial intelligence involve a new type of device? For some American startups, there is no doubt about it. Humane failed his first attempt with his video projector which is worn on the t-shirt, Rabbit hopes to do better with his sort of square Tamagotchi, the R1, which responds to the voice. The brand is banking on a reasonable price of $199 to make it attractive, but its interest remains to be shown.
Advertisement
Unfortunately for Rabbit, the bet seems to be going badly. In addition to mixed first tests and numerous bugs, Rabbit is currently the victim of a leak of its source code. We discover a not-so-original product, which simply transforms voice commands into web requests.
A very artificial artificial intelligence
If Rabbit's source code was leaked, it's because someone wants to harm the company. The author of the leak, who published the code on Github (it has since been deleted), is called “rabbitscam”, which could be translated as “the Rabbit scam”. Its objective is to prove that the hype created by the brand is artificial, since its product really has nothing original. Note that Jesse Lyu, the creator of Rabbit, speaks of a “ rumor ”, does not confirm its authenticity and refuses to talk about source code leaks.
According to Rabbit, the rabbit r1 is a connected object designed to interact with a large multimodal language model. He can answer questions, but also analyze images, thanks to his camera. The brand also highlights an integration of Spotify or Uber, even if these services only partially work for the moment. Several users are complaining about bugs that prevent them from accessing their music or food orders.
Does the rabbit r1 natively integrate a Spotify application? Some thought so, since it is, basically, a mini Android smartphone (it uses a mid-range chip). However, according to the leaked source code, Rabbit is cheating. He uses the tool Playwright to go to spotify.com, search and play music. Same thing for Doordash, Uber and Midjourney, for the generation of images… This is why the service works poorly: the slightest error can cause a failure.
Advertisement
The source code is not complete, but suggests that Rabbit is potentially cheating elsewhere, notably on the AI part. The connection with a PC, which goes through a website, is not secure. It could expose the owner of a Rabbit to the risk of hacking.
The smartphone, the big winner of the Humane/Rabbit sequence
At $199 and without a subscription, the Rabbit R1 is less of a UFO than the Humane Ai Pin, which costs three times as much with a mandatory monthly subscription. Still, it is a device with an unclear purpose, almost all of the functions of which already exist on smartphones. The testers seem to rather appreciate its design, which is, admittedly, successful, but is that enough?
Once again, the rabbit r1 proves the technical superiority… of the smartphone. Generative AI seems destined to explode on Android and iOS, rather than on small gadgets of limited interest. It remains to be seen how much Rabbit lied about the features of its product.