Thursday, March 27, 2008

moore and complexity

Sometimes the impact of Moore's Law can surprise you, even if you work on cutting edge technology and should know better. Take the Rubik's Cube.
First, a confession. I bought the book in order to solve the Rubik's Cube. I had several friends who solved it without the book, including one artist who solved it the first time she picked it up -- bloody savants -- but I bought the book. Fine, sue me. Anyway, for many in my generation, the Rubik's Cube will forever be imprinted as an incredibly complex artifact.

But now Moore has caught up. First, mathematicians at Northeastern used 7 terrabytes of storage to prove it could always be solved in less than 26 or fewer moves. Next, a Stanford mathematician has knocked it down to 25 with only 1500 hours of CPU time.

And these kind of attacks are nothing compared to what has accomplished.

But what I found useful about the Rubik's Cube is that it helps demolish some incorrect assumptions about scale and complexity. As inexorable, exponential progress moves more computing and connectivity into everything around us, the set of what is "solvable" is going to expand with it.

And expand in ways that we often won't think to look.

No comments: