This session is going to discuss our research in the area of automated malware clustering and (recently) automated generation of large quantities of different byte signatures from real-world backdoors and rootkits. Through automated graph-theoretical methods, code similarities between superficially different pieces of malicious software can be identified. Furthermore, these algorithms can be extended to extract "stable cores" for an entire family of malware -- portions of code that are, in some form or the other, present in all executables in a family. From these stable cores, other algorithms can generate a large quantity of different byte signatures. These can be used with amusing effects - from mutating byte signatures to performing a bsearch on AV users to identify the malware authors, a lot of applications will be discussed. While this talk won't prevent you from getting owned, it might make it a little bit more complicated to re-own you with the same infrastructure.