Meta is working on new smart glasses - codenamed "Aperol" and "Bellini" - that go way beyond just taking photos.
They’re adding a “super-sensing” mode powered by AI.
When you say “Hey Meta, start live AI,” the glasses switch into real-time facial recognition and environmental awareness.
In plain English?
The glasses could soon recognize people’s faces, tell you their name, and remind you of things based on what you see.
Think: “You’re talking to Sarah - don’t forget her birthday is tomorrow.”
What the Company Is Saying
Meta is testing this “Live AI” feature in current Ray-Ban glasses, though it’s limited to 30 minutes of use for now (battery life is the issue).
They claim the goal is helpfulness: identifying people you know, remembering your to-do list, or giving suggestions based on what’s around you.
They’ve also said they’re “revising their safety and privacy process” - but the full details? Still fuzzy.
What That Means (In Human Words)
Meta’s not just making glasses.
They’re building a wearable AI assistant that sees what you see - and reacts in real time.
It’s facial recognition on your face.
Yes, it sounds like sci-fi.
Yes, it also sounds like a privacy grenade waiting to go off.
If this works, it changes how we interact with people - and how people interact with us.
And if it gets misused, it changes how safe we feel in public.
🔚 Bottom Line
-
Availability: The “Live AI” mode is in testing now, launching with new models expected in 2026.
-
Cost: Unknown - current Ray-Ban Meta glasses start around $300.
Frozen Light Team Perspective
This is one of those moments where we can’t decide if we’re impressed, nervous, or both.
The idea of real-time, on-the-go context is powerful - if it's used well.
But when you put facial recognition behind sunglasses, you’re not just building tech.
You’re messing with social boundaries.
The future of AI wearables will be decided not just by what’s possible, but by what we’re willing to accept.
And Meta knows it.
The bigger story here?
This isn’t about glasses.
It’s about who gets to teach AI what “helpful” looks like - and whether we’ll even notice when the machine starts looking back.
