They can kick our ass at games of Go, poker, and (most of the time) chess. But there are one or two occasions where artificial intelligence can never match up to human intelligence.
The world-renowned mathematical physicist Sir Roger Penrose has drawn up a chess puzzle to demonstrate just that. The puzzle also highlights the nature of human understanding and shows just how different it is to a computer's. The puzzle has been released in celebration of the Penrose Institute, a new scientific institution that hopes to investigate “human brain, creativity, and the interplay between quantum mechanics and general relativity.”
All you have to do is try to defeat or draw against or the computer while playing as whites from the set position (below). Believe it or not, it is possible for you to checkmate the computer. You can play an online simulation of the chess puzzle right here.
The Penrose Institute
“A human looking at it for a short while will ‘see’ what white must – and more particularly, must not – do, and use very little energy to decide this,” James Tagg, from the Penrose Institue, explained. “But, for a computer, the puzzle requires an enormous number of calculations, far too many for even today’s supercomputers."
They do offer some tips: “For the humans who tackle this problem, I suggest you find some peace and quiet and notice how the solution comes to you. Was there a flash of insight? Did you need to leave the puzzle for a while and come back to it?”
Essentially, the computers will always assume it can win. On the other hand, humans will more readily accept and realize a stalemate is fairly easy to sustain.
It doesn’t matter whether you're a human or a machine, they're inviting everyone to try to crack the puzzle – especially machines. Most chess computers are essentially cheaters. After using algorithms to plot the main part of the game, many computers will attempt to finish it by simply having a lookup table (a book), which contains all the end moves and instructions on how to play every position. Because it looks like an “impossible position,” the computer assumes it can’t be found in the lookup book.
The Penrose Institute said that if you’ve developed an AI chess computer, then run its algorithm to see if it can defeat the puzzle without an endgame book. If it does manage it, they would be pretty surprised but interested to know how it found the solution and how much computational power it required.
If you or a computer manages to crack it then drop the Institute a message with, of course, your proof.