Next-Gen Wearables
I design, build, and test spatial experiences across wearable platforms, from AR on Snap Spectacles to AI-driven smart glasses.
Over a Decade in Wearables
Always Had a Gadget On
It started with a Casio watch at six years old and never stopped: over a decade of strapping on whatever's next and figuring out what it can do.
Early Adopter by Nature
When new tech drops, I order it and build with it: get the hardware, break it open, find what nobody else has tried.
The Timeline
- → First Casio watch - 6 years old
- → DIY POV camera rigs - before glasses had cameras
- → Pebble smartwatch - early 2010s
- → Early VR viewers & Oculus DK1
- → Snap Spectacles, Even Realities G1, Mentra Live - now
Snap Spectacles
Spatial AR Development
Spatial Creativity Unlocked
Spectacles bring 6-DOF spatial tracking, hand tracking, and world-locked content. Each Lens Studio experiment explores a different axis: generative AI overlays, real-time 3D rendering, and interfaces that respond to gesture and gaze.
Prototyping in AR
Each session is a rapid prototype: 3D objects anchored to real surfaces, generative image pipelines rendering straight into spatial view.
First Look - Unboxing the Spectacles
First reactions to spatial AR, hand tracking, and world-locked content straight out of the box.
AR Development Lab
Generative AI, OpenAI Realtime, and Snap3D spatial rendering through Lens Studio on the Spectacles dev kit.
Digital Nature
World-locked AR flora growing from the palm, a hand-tracked study in spatial anchoring.
Citi Field - Live AR
Spatial overlays anchored to Citi Field's architecture in real time: large-scale environment tracking during a live event.
Even Realities G1
AI-Powered Smart Glasses in the Field
Japan - Live Translation
Even Realities G1 in Japan: real-time translation in the lens through restaurants, train stations, and local shops.
Breaking the Language Barrier
In Japan, where I don't speak the language, the G1 put live translation in my field of view without ever pulling out a phone.
I ordered meals from translated menus, read kanji subway signs that resolved into English, and had spontaneous conversations with locals while staying present.
The Invisible Interface
The G1 looks like regular glasses, so I wore them everywhere: izakayas, train platforms, Shibuya at night.
The G1 didn't add a layer between me and Japan; it removed one.
OTTO - AI for Smart Glasses
Mentra Live / SPDR Agent Network
O.T.T.O.
Operational Technician & Tactical Observer - a custom AI assistant built for the Mentra Live smart glasses. Voice-first and vision-capable, it observes, remembers, and assists from your peripheral.
Multi-Model Intelligence
OTTO routes through OpenRouter to Gemini 2.0 Flash, GPT-4o, Claude Sonnet, and Llama 3.3, switching models by voice command for vision, creative work, code, or fast inference.
Persistent Memory
Backed by Supabase, OTTO keeps facts, visual observations, work logs, and meeting transcripts across sessions. Ask “where did I leave my phone?” and it recalls from past observations.
Spatial Awareness
The camera adds OCR for signs and menus, live translation across 20+ languages, and a background visual memory.
Core Capabilities
Voice-First Interaction
Wake word activation, 10-turn conversation memory, and 1-2 sentence responses built for the glasses speaker.
Work Mode
Tracks tasks with periodic frame capture, flags off-task drift, and writes productivity summaries with visual documentation.
Meeting Mode
Timestamped transcription, real-time summaries, and automatic extraction of topics, decisions, and action items.
Live Translation
Point-and-read OCR with instant translation across 20+ languages: Japanese menus, Korean street signs, French documents.
Visual Memory
A photo every 2 minutes builds a passive, GPS-tagged visual log for spatial recall: 'what was that store I passed earlier?'
Agent Delegation
Hands research, scheduling, or code tasks to specialized SPDR network agents while you stay in the physical world.
System Architecture
TypeScript Core
Express.js v5 with the Mentra SDK, running as a session-aware server for multiple glasses, with reasoning, memory, and voice commands in separate modules.
Real-Time Dashboard
A web dashboard shows active sessions, current model, recent transcriptions, battery levels, work frame captures, and meeting status, with model switching from any browser.
Hardware Integration
Mentra Live hardware hooks: multicolor LED status, spatial audio cues, camera for vision tasks, GPS for location features, and VAD (voice activity detection) for natural conversation.
SPDR Network
A node in the SPDR agent network: delegates research, code generation, and scheduling to specialized agents and tracks completion.
The Vision
Beyond the Screen
Spectacles, G1, and Mentra Live all point to the same future: computing that lives in the world, not on a rectangle in your pocket.
Always Learning
Every device and field test adds to a working knowledge of how people interact with spatial computing.
