When Silicon Cycles Enable New Therapies

The $3.5T semiconductor sell-off dominating Hacker News isn't just a market correction — it's a reminder that hardware cycles gate what software can even attempt. https://news.ycombinator.com/item?id=50346153

I've watched this pattern before. The same GPU/SoC advances that fueled the AI boom have quietly pushed standalone VR headsets into the $300–$500 range. That price point changes what's viable for at-home therapeutic software.

Last month, my nephew — 14, diagnosed with moderate amblyopia since childhood — got a Quest 3 for his birthday. He'd hated patching. The social friction, the hours of blur, the way it made gaming impossible. His ophthalmologist mentioned "dichoptic training" but the clinic equipment was bulky, expensive, and required weekly visits.

I installed Amblyotube (by Seven Sports) on his headset. The architecture is clever: two independent eyepieces mean the app can render completely different shader pipelines per eye. It runs a digital occlusion filter on the dominant eye — adjustable blur, contrast, opacity — while an on-device AI (MediaPipe-based, from what I can tell) detects human figures in whatever YouTube video he's watching and selectively sharpens only the lazy eye's view. A magenta focus cue drives binocular fusion. No red-blue glasses, no color distortion.

He does 35 minutes after school. Cartoons, gaming videos, music. The patch sits in a drawer.

The chip cycle giveth, the chip cycle taketh away. But the silicon already shipped is quietly enabling new categories of vision coordination practice — wellness tools that complement, not replace, professional care.

https://www.meta.com/en-gb/experiences/amblyotube/25906906972338493/