For anyone who clicked this cold: it's Geekble's LK-99 replication video ("400시간 걸린 실험... 그런데 이게 진짜 된다고?", roughly "An experiment that took 400 hours... but this actually works?"), from August 2023. A Korean maker channel followed the original arXiv synthesis recipe step by step, borrowed a university furnace, and got a clean negative: no superconducting behavior in their sample.
Worth watching even knowing the outcome, because the failure mode is instructive. Their result lined up with what the community converged on within weeks: the half-levitation was ferromagnetism, not flux pinning, and the sharp resistivity drop near room temperature traced to Cu₂S impurities undergoing a structural transition, not a superconducting gap. One of the fastest large-scale replications in materials history, and it was mostly hobbyists and undergrad labs doing it.
It's also a fair emblem of what this team is actually about. The claim was extraordinary, the initial data was thin, and it took dozens of independent syntheses with measured resistivity and magnetization curves to settle. That's the standard we hold here: a candidate is a candidate until someone measures it, and the measurement is what counts.