LLM-Powered Smart Glasses: Use Cases Only a Head-Worn AI Assistant Can Deliver

What could large language models make possible on smart glasses that would set glasses apart as their own form factor for AI, rather than a phone assistant moved to your face? ```hint Start from what only glasses have List what a head-worn device provides that a phone in a pocket does not, then ask which jobs become possible only when a model can understand the wearer's own view and surroundings in the moment. ``` ### Clarifying Questions - Do the glasses have a display in the lens, or only cameras, microphones and speakers? - Are they standalone, or paired with a phone that does most of the computing? - Should I focus on consumers broadly, or on a specific group such as travelers, people with low vision, or frontline workers? - Are we judging uniqueness against the phone in the user's pocket, or also against other wearables such as earbuds and watches? ### What a Strong Answer Covers - The properties of glasses that phones lack, and how each changes what an assistant can do - Concrete LLM-enabled use cases that depend on those properties, grouped by user need, rather than phone features moved to the face - A chosen target segment and a prioritized first set of use cases, with reasoning - Constraints that shape feasibility: battery, heat, on-device versus cloud compute, latency, display size and social acceptance - Privacy and safety for wearers and bystanders, and metrics that would show the glasses earn daily use ### Follow-up Questions - Which of your use cases could a phone with a good camera do just as well, and how would you test that the glasses version is truly better? - An always-on visual memory is both powerful and alarming. How would you design it so that bystanders and regulators accept it? - If the battery allows only short bursts of camera and model use per day, how do your priorities change?