From Slashdot: a dude named Thomas Rocicki has just proven that 23 moves is sufficient to solve an arbitrary Rubik’s cube configuration, not including the impossible configurations created by dismantling and re-assembling the cube, the only method by which I was able to reliably solve it.
Rocicki had previously established it could be done in no more than 25 moves, but access to more compute power allowed him to shave two moves from the his cube-solver algorithm. Pretty cool. I’m sure if the Franklin Ace 1000 (props to you, Howard!) I used in my youth had the power of the renderfarm at Sony Pictures Imageworks, I would have been able to prove this using my highly sophisticated BASIC skills. Yeah, right.
Now this guy’s algorithm just needs to be connected to the Lego Mindstorms Rubik’s cube-solving robot I heard about years ago, which is also an incredible piece of engineering. Watch the video below to see it in action…