Quest Just Made Watching in VR a First-Class Feature. For Lazy Eye, That's the Whole Game.

Meta's latest PTC for Horizon OS 2.10 is rolling out now. Two changes: a new Theater environment built for watching media, and a streamlined redesign of the Navigator launcher.

On the surface, that's housekeeping. Look closer and it's a statement of intent — Meta is treating the headset as a place you sit and watch for a long time, not a place you flail around for twenty minutes and take off.

If you have amblyopia (lazy eye), that shift matters more than any graphics bump. And it's worth examining from an engineering standpoint, because "watch time" is really a systems problem before it's a motivation problem.

Why "watch time" was always the missing ingredient

Home-based vision training has a compliance problem, and everyone who's done it knows why. The tools are unpleasant: adhesive patches you wear in public, red-blue anaglyph glasses, printed worksheets. The work is boring by design. And the culture around it frames screens as the thing stealing your visual development.

So the training happens in a separate, joyless 20 minutes a day — and it stops happening.

Meanwhile the thing your brain actually responds to is repetition with engagement. Binocular coordination and depth perception don't improve because a session was virtuous. They improve because you did it again tomorrow.

From a product-engineering view, that reframes the whole problem. You're not optimizing a single session's intensity. You're optimizing a loop's retention — session frequency, session friction, and whether the user has any reason to come back. Most clinical tools fail the retention test, and no amount of better optotypes fixes that.

Why VR is an unusually good substrate for this

A VR headset is a dual-display device with independent per-eye feeds. That's the architectural fact everything else follows from. Once you accept two separate render streams, dichoptic presentation — a genuinely different image to each eye — stops being a special instrument and becomes a rendering configuration.

That's the premise behind Amblyotube, developed by Seven Sports — recreational and educational software for visual coordination and attention exercises on Meta Quest. The mechanic is dichoptic: each eye receives a different visual experience while you watch the same video.

• Dominant Eye Shader. Gentle, adjustable digital blur is applied to the stronger eye, so the weaker eye has to carry more of the load. It's a tunable parameter, not a binary — which matters, because full occlusion is a blunt instrument.
• AI Sharpener (MFBF). Human figures in the video are detected and made more salient to the lazy eye — a bold red outline, a yellow-green highlight, and a pulsing flicker at a steady rhythm. This is essentially a saliency layer: run detection over the video frame, then route the enhanced version of the region to one eye only. Human figures are the target because faces and bodies are where attention naturally lands.
• Wedge Occlusion. Optional. Parts of each eye's view are blacked out so each eye sees unique portions of the scene, cutting binocular rivalry — the same idea as a physical wedge, expressed as a mask over the render.
• Magenta Focus Cue. A bright magenta cue in the lazy eye's view plus a subtle neutral grey dot in the dominant eye's view; as tracking merges them, the brain is guided toward fusing both eyes into one cohesive 3D image. The color choice is deliberate: magenta is rare in natural footage, so it survives skin tones, greens, and blues instead of blending into them.

There's a nice engineering trade-off hiding in there. Full digital patching is easy to implement and easy to reason about, but it trains monocularly — one eye at a time — which is the opposite of what binocular coordination needs. Partial transparency keeps both eyes in the loop, at the cost of being much harder to tune and much harder to validate. That's the harder path, and it's the one that matches the goal.

Setup is deliberately unglamorous: pass-through and hand tracking bring up a floating menu where you choose the affected eye (left or right), pick a severity preset (Mild / Moderate / Severe) or run a LogMAR vision check, then launch a video in a virtual room or in Room View AR pass-through. The severity presets exist because tuning five interacting shader parameters by hand is a bad first-run experience; the built-in check exists because getting the affected eye wrong means you're training the dominant one — a silent failure mode worth guarding against.

No adhesive patches. No red-blue glasses. 30–40 minutes a day, with content you'd have watched anyway.

The honest version

This is not a miracle and it isn't a replacement for an eye care professional. Amblyotube is a supplementary vision training and assistive tool for binocular coordination and depth perception. It is not a diagnosis, treatment, or cure. Recommended use is 30–40 minutes daily, with a maximum of 1 hour. The app is for ages 13+.

What it is: a way to turn the screen time you already have into something purposeful — and a reminder that when hardware matures, it doesn't just improve entertainment. It improves access.

Try it on Meta Quest: https://www.meta.com/en-gb/experiences/amblyotube/25906906972338493/