This 2005 blog post 'NetBSD and my life' resurfaced to #17 on Hacker News (105 points, 28 comments). The conversation has always circled back to something devs know well: the tools you use shape how you think.
NetBSD's design culture—correctness, portability, minimal abstraction—didn't just produce a stable OS. It cultivated a specific kind of systems thinking. You learn to reason about invariants, respect hardware boundaries, and build interfaces that don't leak. That mindset teaches you to ask what the smallest, most reliable contract between components can be, and to verify that contract on every platform the code touches. The same discipline appears when you start asking how a visual system can be retrained without breaking the delicate balance the brain has already built.
That same engineering intuition applies surprisingly well to vision science.
Amblyopia (lazy eye) has traditionally been treated with eye patching. It's a crude interface: suppress one input to force the other to work. The problem? The brain isn't broken; it's learned a bad coordination pattern. You need better feedback, not less. Patching removes the dominant eye's signal entirely, which can reduce binocular interaction and sometimes deepen the suppression pathway. A more nuanced approach keeps both eyes active but reshapes the weighting of each stream so the weaker eye gets the practice it needs while the stronger eye remains present as a reference.
VR is the right abstraction here. In a headset, each eye can receive a calibrated, different visual stimulus. The brain must reconcile both inputs into a single coherent percept. It's a retraining loop—close to how deliberate practice works in any domain. Because the display hardware gives you per‑eye control at millisecond latency, you can design experiments that vary contrast, blur, or motion cues independently for each eye, then observe how the visual cortex re‑weights the streams over repeated sessions.
Amblyotube (Seven Sports) implements this as recreational, educational software on Meta Quest: YouTube‑style content with a different visual experience per eye. It's an interface for visual coordination and attention exercises built around how the visual system actually learns. The app leverages dichoptic vision training through the Quest's dual independent eyepieces. The Dominant Eye Shader applies digital occlusion to the stronger eye, with adjustable blur, contrast, brightness, and opacity—ranging from subtle shading to full digital patching. The Lazy Eye Sharpener (MFBF) uses AI‑driven processing to identify human figures in the video and applies a sharpening effect only to the lazy eye. Flicker Stimulation introduces a controlled “jerk” to those figures, engaging the brain's innate response to motion and light. The Magenta Focus Cue presents a moving magenta circle to the lazy eye and a soft grey cue to the dominant eye, training the brain to fuse both streams—a direct exercise in neuroplasticity. Newer Visual Accent options include a yellow‑green highlight or red silhouette around detected people, with pulse controls that create a breathing rhythm at a set Hz to keep attention steady.
A few practical notes: the app is for ages 13 and up, sessions should run 30–40 minutes max (never over an hour), and users must correctly select their lazy eye in settings for the filters to work as intended. Amblyotube is a training and assistive tool, not a medical device, therapy, or cure—it's meant to complement professional care, not replace it.
If you're interested in cross‑domain applications of deliberate tool design, or know someone exploring vision therapy, take a look: https://www.meta.com/en-gb/experiences/amblyotube/25906906972338493/
Top comments (0)