There was an interesting paper just uploaded to the cryptology e-print archive entitled: Near Collisions for the Compress Function of Hamsi-256 Found by Genetic Algorithm. The function — Hamsi 256 — is one of the candidates for the NIST hash competition. I am not sure how damaging this is for the Hamsi hash function, but this paper piqued my interest since I am in general somewhat cynical about the effectiveness of genetic algorithms and so it was cool to see them in action.