Bee Here Now

I’m not sure people really understand what is coming with AI and wearables. I say that, in part, because it’s already here.

A couple weeks back, Amazon (re)launched the Bee AI device with some fanfare, but not much. Certainly not as much as there should have been. Because the device is… pretty amazing? Definitely for a device that costs a mere $75.

I had been tracking Bee since shortly before Amazon acquired it because I was interested in AI wearables but in particular, super simple AI wearables. To me, Humane clearly crapped the bed simply because it was roughly ten times too complicated — and yes, overhyped. The world wasn’t ready for laser projections on your hands. I mean, I’m not sure the world will ever be ready for that, but I digress… The Rabbit device gave good demo, but little else. The Friend pendant thing was closer, but also too cute (while also perhaps being too creepy). Limitless was perhaps the most promising (perhaps why Meta quickly acquired it). What we clearly needed was a super simple wearable, perhaps with just a single button to invoke a chatbot.

That’s basically what the Bee device is. A wearable with one single button to rule them all. Only it’s even more simple than that. Because it’s not even a wearable you talk to, it’s just one that listens to you. And everyone you interact with. And then gives you a summary of what was discussed, and suggests some follow-up ideas and actions based off of what it hears.

But the really crazy thing is just how good it is. I bought one almost on a whim because again, I’ve been watching this space and it’s $75 on Amazon. I’m honestly shocked just how good it is at what it does. There’s some lightweight onboarding to be able to recognize your voice, but once you do that, it’s off to the races.

The Bee quickly was able to deduce my various family members — even though I didn’t explicitly input their voices — including their names and ultimately, their wants and needs. This sounds a bit creepy because it is a bit creepy. And it’s going to creep the fuck out of some people. But it’s also sort of awesome just how far this technology has come in just a few short years.

And so I’m torn. I’m not sure if this is a device that every single Big Tech and Big AI company is going to copy ASAP, or if it’s going to garner an immediate backlash and be banned everywhere.

The latter bit may be why Amazon seems to be underplaying just how big of a deal this device is. It’s a big deal because it’s so good at what it does. And what it does is listen to all of your conversations.

Amazon, of course, says all the right things about privacy and data protection, but that may not end up mattering as much as the societal pushback here. Meta is already feeling it with the Ray-Ban smart glasses. And even Apple, the most trusted brand in tech, has gotten some blowback against products that leverage this type of surveillance. And that’s before such products and features are even live yet!

Recording capabilities or not, people clearly don’t like the notion of wearables that can be always-on and always listening or watching. Hence, the “PervPods” nickname. And actually, it sure sounds like an upcoming feature of the Apple Watch is very similar to what the Bee offers.

Meanwhile, beyond their Ray-Ban smart glasses, Meta has their forthcoming Muse Charm device which could offer some similar functionality to the Bee. And then there’s OpenAI’s forthcoming device made in conjunction with Jony Ive and his team. These two devices sound more robust — like actual AI agentic devices beyond simple listening tech. Still, they’re all clearly dancing around similar ideas.

And at the highest level, that idea is: why type into an AI chatbot when it can just always be listening and figuring out what you need before you explicitly ask for it? Given just how good Amazon is at this with Bee, you have to wonder how killer of a product an Anthropic wristband would be? And an OpenAI one. And a Meta one.

Again, a super simple wearable with a single button to turn its capabilities on and off. No screen. No laser projectors. Hell, not even a speaker to output responses back to you. Those can be found in your app. The key is the input and the backend smarts to do something interesting with it. It’s sort of mind-blowing.

Here’s a wild, fun real world use case: I actually had a hotel dispute this week where they charged me for a drink at the bar that I didn’t order. The Bee caught the entire conversation with the bartender and confirmed my recollection. It wasn’t a recording — though it can do that too — but it was a summary of the interaction. Is that legal? I mean, depending on the state, certainly not in a court of law. Is it interesting? Obviously! Can it be manipulated in ways? Probably!

Again, I just can’t believe how well it works. It ingests, does some processing, and the summaries and bullet points (and suggested action items) are so good. Maybe not 100% accurate, but close enough for the purposes of aiding recall.

And given the Amazon ownership, and price point, it feels like the start of a new state of surveillance tech. That sounds creepy, and it is in some ways. But it also feels inevitable. The question now is if this early tech is embraced or shunned. It could go either way…