The hardest part here was the diagnosis. The maker burned hours debugging before finding that the power supply itself was the culprit, and the way it was finally proven is neat: the ESP32 controller is lifted by hand and held over the supply, and when it gets close the strip freezes, glitches, or cuts out entirely. Move it away and the lights come straight back. That is a clean, repeatable test for a fault that must have looked random for a long time.
The build itself looks tidy. It is an ESP32 running WLED driving an addressable strip, housed in a white plastic enclosure with blue terminal blocks. Power comes from a metal switching supply with a fan intake, joined to the DC cabling through a yellow XT60 connector. The mains side is properly dressed, with brown and blue wires plus a yellow and green ground. The strip runs along an overhead surface, showing a moving rainbow effect whenever the controller is kept clear of the supply.
The details that stand out to me are the hands-on test and the fact that the maker shared it as a warning for anyone chasing similar issues. Watching a controller misbehave on demand, just by bringing it near the supply, makes the cause unmistakable.
The open question is the fix. No permanent solution to the interference has been stated, and the strip type and length were never given either. The post text was left empty, so the failure details only exist in the demonstration itself. I would love to see a follow-up once the controller has a settled home away from the supply.