How, then, do we evaluate malware similarity and "newness" in a way that both scales to tens of millions of samples and is resilient to the zoo of obfuscation methods that malware authors employ? However, she has continue to remained reasonably desirable ever since then, depending on the artist.