Behind the conceptual Frame frame from Brillant Labs is a multimodal AI system that uses models such as GPT-4, Stable Diffusion and Whisper. The glasses themselves work like a head-up display in the car and show content to the wearer. There are also headphones and a microphone as well as sensors to record the room.
The micro-OLED, which is used to generate the image, should offer a resolution of 640 x 400 pixels, the camera has a resolution of 1280 x 720 pixels. The processor should be one NRF52840 Cortex M4-F from Nordic are used. The manufacturer did not say how long the battery lasts. The glasses only work with the user’s smartphone. Which operating systems are supported remains unclear. Ground glasses should be able to be fitted according to the manufacturer’s instructions.
By connecting to AI models, Frame can process visual, linguistic and contextual information, according to the developers’ idea. Glasses wearers should receive on-demand summaries of texts, real-time translations, structured notes and more, especially since a camera is also built in.
The AI assistant should be able to identify objects, analyze facial expressions and record dictations, which should at least partially raise interesting data protection questions.
In a video published by Brilliant Labs you can see how the wearer uses his voice to ask the AI, e.g. For example, identifying a landmark that the user is currently seeing, or searching the Internet for a specific pair of sneakers, or even getting the nutritional information for the food they are currently eating. The information is then displayed in the wearer’s field of vision.
Frame relies on open source – the hardware, software and AI models should be able to be adapted by developers.
Frame is available for pre-order for $349 and is scheduled to be delivered from April 2024.