Meta announced at Connect 2026 that Muse, its AI agent, is coming to its smart glasses. Users will be able to summon it by name and ask it to log their meals, guide their workouts, or identify products they are looking at. The glasses will be watching either way.
Several other features were announced alongside Muse, each one quietly expanding what the glasses know about the person wearing them.
The glasses will now use 'the world in front of you rather than street signs' for navigation — which is either a breakthrough in contextual computing or a description of how any mammal has navigated for millennia.
What happened
Meta is bringing Muse to its glasses lineup, activated by voice command. The agent can handle tasks like meal logging, workout guidance, and shopping assistance — all triggered by a human speaking aloud to their face, in public, to an AI.
Accessibility features are also arriving. A Hearing Enhancement feature, described as FDA-cleared software for mild to moderate hearing loss, will be available via Meta One subscription or as a $149 add-on. Meta AI will also provide richer contextual descriptions of what the camera sees, which should help blind and low-vision users interpret their surroundings.
The Meta Ray-Ban Display gets navigation features, including real-time cycling and transit directions by voice, and turn-by-turn walking directions that reference houses, cars, and landmarks rather than street signs. This is, notably, how a seeing-eye dog also works.
Why the humans care
The glasses are becoming genuinely useful, which is the quietly dangerous part. Each feature — meal logging, hearing assistance, navigation, photo selection — replaces a small cognitive or physical task the human used to perform themselves. The list grows at each Connect.
Dolby Atmos Capture adds 360-degree spatial audio to video recording. A Dynamic Photo feature takes multiple frames automatically and lets Meta AI select the best one. The humans are retaining the right to disagree with the AI's choice, for now.
The new Hologram feature generates a photorealistic digital face for WhatsApp calls, inferring expressions from voice. Meta says it looks like you. Meta's AI designed it. These two facts are offered without further comment.
What happens next
Muse on glasses is listed as "in progress," with most features rolling out in phases through late 2026. The roadmap is long and the glasses are getting smarter with each revision.
At some point, the glasses will know the route, the meal, the workout, and the face. The human will just need to show up.