The Sunset Date Is on the Headset. Not on the Habit.

Meta is winding down Quest 2 and Quest Pro support, with a final OS update early next year and HorizonOS development being pruned ahead of Meta VR Glasses next spring. That headline reads like a hardware story. For a specific group of users, it's a routine story — because what you can actually do inside the headset outlives the headset.

Here's the uncomfortable part of that news. If your daily visual training lives on a device that just got a sunset date, your habit now has an expiry date attached to it. Amblyopia and lazy eye training is one of the most habit-dependent things there is: the science of it doesn't reward dramatic one-off sessions, it rewards showing up 30-40 minutes a day, over months, with something you don't dread.

That's the gap Amblyotube, developed by Seven Sports, is built for. It's recreational and educational software for visual coordination and attention exercises in VR, and it runs on Meta Quest. The core idea is simple: instead of a clinical exercise, you watch YouTube-style content — the kind of thing you'd already have on — while the app runs a different visual experience for each eye.

I want to walk through how that actually works, because the engineering here is more interesting than the marketing copy suggests. This is a real-time stereo rendering problem with a therapeutic constraint: each eye needs to receive a deliberately different image, and the difference has to stay stable frame to frame while video plays at full rate. You can't just slap a filter on one eye and call it a day — the two views have to remain fusable enough that the brain keeps trying, but divergent enough that the weaker eye is forced to do real work.

What that looks like in practice:

The dominant eye gets a gentle, adjustable digital blur (the Dominant Eye Shader). The amblyopic eye gets an AI Sharpener (MFBF) that detects human figures and marks them with a bold red outline, a yellow-green highlight, and a pulsing flicker at a steady rhythm. Optionally, wedge occlusion blacks out part of each eye's view so each eye sees unique portions of the scene — which cuts down on binocular rivalry instead of letting the stronger eye quietly dominate. And to pull the two images together, a bright magenta focus cue appears in the lazy eye's view with a subtle neutral grey dot in the dominant eye's view; as you track them, they merge into one cohesive 3D image. That last piece is the point — it's training the brain to fuse both eyes and process real depth.

Let me unpack the trade-offs, because they're where the design gets opinionated.

Why asymmetric rendering, not a global filter. A naive approach would be to blur the whole scene for the dominant eye. That's cheap, but it also degrades the content the user actually came to watch, and it doesn't localize the training signal. Amblyotube instead layers the treatment on top of the video stream, per-eye, so the content stays watchable and the therapeutic difference is applied selectively. The MFBF sharpener runs human-figure detection so the red outline and pulsing flicker land on subjects — faces, bodies — which is where binocular fusion matters most in natural viewing. That's a detection pass feeding a per-frame compositor, not a static overlay.

Why the flicker is rhythmic and steady. The pulsing flicker on the highlighted figures isn't decoration. A consistent temporal rhythm gives the amblyopic eye a predictable target to lock onto, which matters when that eye is the one being asked to lead. Predictability beats intensity here.

Why wedge occlusion is optional. Full-field viewing lets the strong eye dominate. Wedge occlusion blacks out part of each eye's view so each eye sees unique portions of the scene — this is a rivalry-management technique, and making it optional lets users dial the challenge up or down instead of committing to one rigid mode.

Why the focus cue is two different colors. The magenta cue in the lazy eye's view and the neutral grey dot in the dominant eye's view are a fusion pair. The user tracks them; as the two converge, the brain merges them into a single 3D image. It's a guided convergence exercise disguised as a game mechanic.

Setup is on your own terms: pass-through and hand tracking on a floating menu, pick the affected eye (Left/Right), choose a severity preset (Mild/Moderate/Severe) or run the LogMAR vision check, then launch a video in the virtual room or in Room View AR passthrough. The severity presets matter because the blur and sharpener intensities need to scale — a Mild setting and a Severe setting are doing very different amounts of work, and letting the user pick (or measure via LogMAR) keeps the challenge in a useful band rather than either trivial or overwhelming.

No adhesive patches. No red-blue glasses. Just a headset and content you like.

Two honest caveats worth repeating. First, this is a daily practice, not a sprint: 30-40 minutes recommended, max 1 hour, ages 13+. Second, and more important — Amblyotube is supplementary vision training and an assistive tool for binocular coordination and depth perception. It is not a diagnosis, a treatment, or a cure. If you have amblyopia, work with an eye care professional on your actual plan.

For builders, the takeaway is broader than one app. The interesting constraint in XR right now isn't resolution or FOV — it's building software whose value survives the hardware cycle. Per-eye rendering pipelines, on-device detection, and fusable stereo overlays are portable techniques. The Quest 2 sunset is a reminder to design habits and tools that outlast the device they shipped on.

The Quest 2 chapter is closing. The habit doesn't have to. https://www.meta.com/en-gb/experiences/amblyotube/25906906972338493/