NIXFollow the research ↗
NUETEO explores a new way to create and communicate through a noninvasive interface built directly into your glasses.
Research vision · Concept hardware and simulated experiencesNUETEO is the intent interface envisioned inside the glasses. Temple-contact sensors, personalized models, and onboard XPU inference would work together to translate deliberate, trained intent into digital actions.
The focus is what you intend to express, create, or do. From a familiar object to a connected idea, the ambition is to make room for a new form of communication.
View a specific apple image during guided calibration. Repeated prompts ask you to attend to or imagine it while temple sensors record synchronized signals. The research objective: reproducible associations between measured patterns and labeled intent.
Intentionally activate a session and recall the trained apple concept. The proposed system attempts to classify the signal against your personal model, applying confidence thresholds and abstaining when uncertain.
An onboard XPU would use the inferred concept to condition a generative AI model. A new apple image would appear in the glasses, shaped by the prediction and your chosen refinements.
Preview the result, refine it, choose a recipient, and confirm transmission. The ambition is creation and sharing through trained intent while keeping every action deliberate.
The AI generates an image from an inferred concept. It does not extract an exact picture from your mind.
A connected idea can be expressed through images, language, sound, spatial scenes, and actions. Open-ended concept decoding is a future research goal.
“Meet me at our usual place
after sunset.”
Shared context: only information both participants have explicitly permitted the system to use.
A proposed semantic representation would link the person, place, time, and invitation. After review, a receiving device could present that meaning as words, audio, a map, or a spatial scene.
The long-term ambition: preserve intended meaning across different forms of expression.
The recipient receives a device-mediated expression, rather than understanding transferred directly into their brain.
Collect signals through noninvasive temple contacts.
Estimate a trained intention using a personalized model.
Map the estimate to a structured concept or command.
Produce an image, message, or spatial representation.
Let the wearer review and authorize the result.
Send the approved representation over conventional networks.
Sensor measurements are inputs. Model predictions are estimates. Generated content is an AI-created expression. Temple sensors alone are not evidence of reliable, unrestricted neural decoding.
A visual introduction to the idea: an image, an inferred concept, and a possible expression. No neural signals are measured here.
A familiar apple makes the proposed relationship tangible: an inferred concept could guide a generative model toward an image.
The AI generates an image from an inferred concept. It does not extract an exact picture from your mind.
Pre-generated artwork and illustrative traces. This is a visual explanation, not a working interface or a demonstration of model training.
Person + place + time + intention → one reviewed invitation.
Assemble an invitation to Alex, using an explicitly permitted familiar place and a time after sunset. Switch how the recipient’s device presents it.
No real recipient, location lookup, audio playback, or message transmission. The same intended message is preserved in every presentation.
Transparent XR glasses with an intent interface envisioned inside. Concept hardware; this render does not establish technical feasibility.
Product design commitments for a proposed interface. Neural and physiological measurements are sensitive personal data.
Intentional activation, a clear sensing indicator, and explicit handling of uncertainty and unintended activation.
Local processing wherever feasible. User control over calibration data and model deletion.
Explicit permission before accessing shared context, with clear boundaries around what the system may use.
Review inferred meaning and generated content. Confirm every transmission before the representation is sent.
NUETEO is a proposed research direction. Reliable noninvasive recognition of rich, open-ended intent from glasses-mounted sensors remains an unresolved challenge. Demonstrations on this page illustrate the intended experience.
This page presents proposed workflows and conceptual hardware, rather than demonstrated performance. Open-ended intent decoding remains a future research goal.
“Communication at the speed of intent” is an aspiration. Sensing, inference, rendering, and network propagation each introduce delay.
NUETEO by NIX — From intent to shared understanding.
This page is the current research vision. Bookmark it to follow future updates. A public research feed and mailing list have not been announced.