Meta tests always-on 'super sensing' mode for next Ray-Bans
TL;DR
- Meta is testing a 'super sensing' mode that keeps Live AI running in the background for hours, up from roughly 30 minutes on current Ray-Ban Meta glasses.
- Mark Zuckerberg is reported to have questioned whether the mandatory white capture LED could stay off during the always-on Live AI mode, and Meta is weighing the idea.
- Two next-generation devices codenamed Aperol, a sunglasses design, and Bellini, a prescription pair, are aimed at late 2026 or early 2027.
Something shifted in the wearable story this week, and it isn't the widely-covered tamper-detection announcement. According to the Financial Times, Meta is testing a mode for its next-generation Ray-Ban glasses in which Live AI runs in the background for 'hours' instead of the roughly 30 minutes the current device supports. Cameras and sensors stay active continuously, so the assistant can remind you to grab your keys on the way out, or nudge you to stop in a store for dinner ingredients. Internally the company calls the feature set 'super sensing,' and it is being built into two devices reportedly codenamed Aperol, a sunglasses design, and Bellini, aimed at prescription wearers.
The reason this is more than a spec bump is the design question sitting inside it. Mark Zuckerberg is reported to have questioned whether the small white LED that currently lights up whenever the glasses are capturing could stay off in Live AI mode, and Meta is said to be weighing the idea. That LED is the entire social contract with anyone standing near a wearer. Turning it off in an always-on mode would flip the default that camera glasses have leaned on to date, in which the person around you was at least in principle notified when the lens was hot. It lands the same week Meta announced that its AI glasses will disable the camera if that same LED is physically blocked, tampered with or destroyed, a policy that reads very differently once you know the company is separately debating switching the light off itself.
The honest caveat is that this is pre-release reporting on hardware Meta has not confirmed, sourced through leaks and unnamed people. What the reporting doesn't tell you is how continuously captured audio and imagery would be processed, where it would live, for how long, or whether the facial recognition capability Meta is 'exploring' for name recall would be regionally gated. Launch timing is also fluid, with leaks currently pointing between late 2026 and early 2027.
If Meta does ship always-on capture without a visible recording light, the strategic opening is for competitors and enterprise buyers who can credibly promise the opposite: opt-in sessions, on-device processing, and an indicator that never goes dark. The most honest upside case for the Meta version, if the privacy piece can be squared, is accessibility, where a Live AI that actually lasts a full outing rather than half an hour is a genuine unlock for blind and low-vision users.
Shared on Bluesky by 1 AI expert
Originally reported by ft.com
Read the original article →Original headline: FT: Meta Testing Ray-Ban Smart Glasses That Continuously Record Audio and Snap Photos Every Few Seconds for Recall-Style AI Memory