Computer scientists develop 'mathematical jigsaw puzzles' to encrypt software (UCLA)

Ultimately, the computer steps from a to b to c to d. A lot of work has gone into compilers to make sure the steps from a to b to c to d are as small as possible, and arranged carefully to match the processor. And this work is claiming that it can find a way to re-rig everything, and still have the attacker not understand what is going on?

I'm suspecting a Denmark-sized piece of limburger cheese - it's just not passing the smell test to me.