No one has yet shown that SHA-1 collisions are possible, as with this MD5 technique, but it's known that SHA-1 isn't as strong against potential attacks as it was designed-to-be.
That knowledge has driven adoption of SHA-2, and was partial motivation for the (completed in 2012) competition to design SHA-3. I don't believe any weaker-than-designed problems have yet been found with the SHA-2 family.
In 2012, Bruce Schneier reported on an analysis by Jesse Walker of Intel about when SHA-1 collisions might be practical to create:
That analysis suggests: "A collision attack is therefore well within the range of what an organized crime syndicate can practically budget by 2018, and a university research project by 2021."
But it also notes non-commodity approaches (GPUs, custom chips, etc) could achieve SHA-1 collisions sooner/cheaper.
That knowledge has driven adoption of SHA-2, and was partial motivation for the (completed in 2012) competition to design SHA-3. I don't believe any weaker-than-designed problems have yet been found with the SHA-2 family.
In 2012, Bruce Schneier reported on an analysis by Jesse Walker of Intel about when SHA-1 collisions might be practical to create:
https://www.schneier.com/blog/archives/2012/10/when_will_we_...
That analysis suggests: "A collision attack is therefore well within the range of what an organized crime syndicate can practically budget by 2018, and a university research project by 2021."
But it also notes non-commodity approaches (GPUs, custom chips, etc) could achieve SHA-1 collisions sooner/cheaper.