This is not new at all, I remember playing with MD5 collisions in 2006, but it's good to be discussed from time to time so that we know we cannot trust such things.
We're getting there with reduced SHA-1[1] (that is, less than 80 rounds, that means less than 2^80 theoretical operations[2]). But the cost of finding a collision decreases over time[3], and this is why everybody says SHA-1 is obsolete.
We're getting there with reduced SHA-1[1] (that is, less than 80 rounds, that means less than 2^80 theoretical operations[2]). But the cost of finding a collision decreases over time[3], and this is why everybody says SHA-1 is obsolete.
[1] http://eprint.iacr.org/2010/413.pdf [2] https://www.schneier.com/blog/archives/2005/02/sha1_broken.h... [3] https://www.schneier.com/blog/archives/2012/10/when_will_we_...