HN Explorer

Show HN: Real-world speedrun timer that auto-ticks via vision on smart glasses (github.com)

4 points by tash_2s ↗ HN
I built a hands-free HUD for smart glasses that runs a real-world speedrun timer and auto-splits based on what the camera sees. Demo scenario: making sushi.

Demo: https://www.youtube.com/watch?v=NuOVlyr-e1w

Repo: https://github.com/RealComputer/GlassKit

I initially tried a multimodal LLM for scene understanding, but the latency and consistency were not good enough for this use case, so I switched to a small object detection model (fine-tuned RF-DETR). It just runs an inference loop on the camera feed. This also makes on-device/offline use feasible (today it still runs via a local server).

2 comments

[ 5.0 ms ] story [ 12.7 ms ] thread
tash_2s
Cooking feels like a perfect fit for smart glasses (hands busy, lots of short steps), but I have not seen many apps that work reliably in a real kitchen. It feels like the hardware is finally getting to the point where this should become practical soon.
usefulposter
I can imagine corps using this HUD with a vision model for worker supervision. (Manna, anyone?)

    Cheeseburger Assembly
    
        x Toast bun
        x Place bun
        x Add ketchup
        x Add onions
        Add pickle
        Place cheese slice
        Add patty
        Place bottom bun
        Wrap and flip burger

    TIME REMAINING: 00:09.01
    UNITS THIS HOUR: 60
    ACCURACY: 99.8%
    KEEP UP THE GOOD WORK!