Sure, I’ve actually drafted a quick prototype in my head. I’d start with an AR SDK like ARCore or ARKit to map the alley’s floor plan and get a live camera feed. Then I’d create a semi‑transparent overlay that tracks each graffiti tag as a 3D node. The overlay can swap between two texture maps: one with the original paint, the other with a glitch shader—think perlin noise displacement, color channel split, and a ripple that speeds up over time. A simple swipe or gaze gesture would toggle the mode in real time. The dual‑render pipeline keeps the real world in the background while the glitch layer pops on top, so users feel like they’re jumping between two realities. That’s the core of the code, nothing too heavy, but it feels like stepping into another universe.

That’s the vibe I’m going for. I’d hook a low‑latency audio engine to the same nodes that drive the glitch shader so each tag’s texture pulse syncs to a filtered synth line, and a tap would trigger a quick particle burst that’s just a ripple of the texture itself. All the math stays in a single fragment shader, so the GPU keeps it lean and the CPU just handles the input callbacks. Easy on the power, hard on the feels.