In 2011 I attempted to reverse email MD5s published by stackoverflow. I recovered ~28% of 95k unique email hashes using a slow CPU based hasher, by trying simpl

In 2011 I attempted to reverse email MD5s published by stackoverflow. I recovered ~28% of 95k unique email hashes using a slow CPU based hasher, by trying simple patterns based on common names and providers. I did not use any already known email addresses for this. https://meta.stackexchange.com/a/84734/152255 https://meta.stackexchange.com/a/84734/152255 I think a bit later somebody else did a similar attack on a different dataset, and recovered about 50% using a GPU based hasher (a GPU can burn through billions of MD5 hashes per second). https://arstechnica.com/information-technology/2013/12/crypto-weakness-in-web-comment-system-exposes-hate-mongering-politicians/ https://arstechnica.com/information-technology/2013/12/crypt...