Meta's Glasses Want to Listen Forever: "Super Sensing" Plan Exposed, and the Light Stays Off
On July 8, the Financial Times dropped a bombshell.
Meta is internally prototyping AI smart glasses codenamed "super sensing." The core capability: always-on audio recording, plus a photo automatically taken every few seconds.
Wear them, and they capture everything you see and hear all day. Then you can ask Meta AI — "where did I leave my keys" or "what did that person say at lunch."
"Glasses evolving from something that answers questions into a personal agent that's with you all day, helping you remember things and achieve goals." — Mark Zuckerberg
The Wildest Part: They Want to Turn Off the Light
Current Ray-Ban Meta glasses light up a white LED on the frame when recording — the only signal telling bystanders "I'm capturing." It's the sole privacy cue for strangers.
But FT reports Meta executives plan to turn that light off in super sensing mode.
Their stated logic: "a constantly lit light makes people around you numb, making it easier to miss recording behavior." That reasoning clashes head-on with the whole point of the light.
A Glaringly Contradictory Contrast
Two days before this scoop, Meta pushed a mandatory firmware update:
If the glasses detect the capture LED has been tampered with, the camera is automatically disabled.
Meta is even pulling ads, banning accounts, and suing people running the gray-market "LED removal service."
Tuesday: the light is so critical that tampering bricks the camera, and we'll sue you.
Thursday: the next product will leave the light off.
This isn't hypocrisy. It's a roadmap revealing the real priority — the LED protects bystanders from other people's glasses, but it doesn't protect them from Meta's product roadmap.
No Need to Wait for New Hardware
Here's the sharper edge: the super sensing feature could roll out via software update to already-sold glasses. Meta doesn't need to wait for the next generation — it could push always-on capture to millions of pairs already in the wild.
Two next-gen devices, codenamed Aperol and Bellini, target late 2026 or early 2027. The entry-level Meta Glasses already start at $299 (launched June 23).
The Metadata Play
One architecture under internal discussion: raw images and audio aren't stored and aren't shown to the user — only metadata is extracted and uploaded to servers for Meta AI to query. Supporters say this reduces privacy risk.
Critics say a model where "you can't see the recording and can't see the raw data" creates surveillance that's harder to detect, not safer.
Regulators Aren't Waiting
New York State's court system bans smart glasses from all courthouses starting July 20 — 1,240 courts, prescription frames included. Wear them in and you surrender them to court officers at the door. Pennsylvania, Hawaii, and Wisconsin courts have followed on individual levels.
In June, two security researchers found a complete facial recognition system baked into Meta's glasses app — code that could have been flipped on with a single update. After the finding went public, Meta deleted it.
Zuckerberg's vision is clear: glasses aren't a tool you prompt for answers, they're an all-day personal agent.
But an "always listening" device paired with a "light turned off" — where that road ends, Meta probably hasn't figured out yet. The one certainty: regulators and the public aren't waiting for them to.