Wearable devices
Are We Witnessing the Era of New Wearable Devices or Just a Fleeting Hype for a Couple of Years?
At CES 2024, the Rabbit R1 was presented - a new device aimed at changing our interaction with smartphones. The concept is straightforward: replace the current milliton-app-based interface with an app-free voice interface.
Every action with the Rabbit R1 is a voice command. The device includes a camera, speakers, a button, and a monitor, primarily for response verification. The button is for making requests: you press it and speak, eliminating the need for saying something like “Hello Rabbit”.
The Rabbit R1 operates on the Large Action Model - the creators' version of a specialized LLM for interacting with external services like Spotify, Uber, etc. They claim to have trained the language model to interact directly with these services. An additional feature is that Rabbit R1 can be further trained for new interactions. During the presentation, they demonstrated training Rabbit to make Midjourney requests in Discord. This strange marvel is priced at $199. Opinions are divided: some see it as a groundbreaking innovation, akin to the 'iPhone moment', while others question its uniqueness compared to existing smartphone voice assistants and why it's a separate device rather than an app. I'm in the latter camp, suspecting that Apple might simply upgrade Siri with similar functionality.
But Rabbit R1 isn't the only wearable device making waves. You've probably heard of the AI Pin - a brooch with built-in AI, featuring a camera and projector for gesture control and voice interaction, projecting information onto your hand.
Another example is Tab, a medallion/pendant AI that constantly listens and records surroundings. You can ask it what was discussed during dinner, for example. Today, Tab raised a $1.9M Seed round.
Meta, in collaboration with Ray Ban, took a slightly different approach. They released glasses with two built-in 1080p cameras, streaming and processing video on remote servers, employing various image recognition models (detection, segmentation, classification, etc.).
I'm immensely intrigued by these devices but am struggling to see their practical application. There's a small chance that one of these mini AI companions might make a significant impact, possibly in niche sectors like healthcare or education. Or they might fizzle out without making a mark. But one thing is clear: they're seen as having potential and are receiving funding for development. Let's wait and see what happens in a couple of years, and if any will be featured at CES 2026.