Often the most fascinating parts of the game aren’t the outcome, but the decisions that lead one there.
Hearing a great commentator discussing a sport about which you know little can be one of the most exciting and illuminating of experiences.
That sensation of being rapidly inducted into strategies which were apparent, but not wholly understandable never fails to intoxicate.
I often feel this when hearing developers talking about videogames. In particular, some of the best demonstrations I’ve seen have seldom been about breathtaking new graphical capabilities, but artificial intelligence (A.I.).
Whilst new rendering technologies providing near photo-real graphics are usually touted as the enablers to making worlds which we can truly believe in, the truth is that the algorithms which determine the behaviour of in-game elements are far more persuasive.
The pursuit of convincing A.I. is a fascinating subject. To understand and create persuasive A.I.
Behaviour, one has to understand at least a little of what it is to be human. Fascinating though it is however, there’s always a slight feeling of discomfort that someone, somewhere is attempting to distill that slight feeling of discomfort into something that can be algorithmically reproduced.
Thinking Machines 4 is a more palatable version of that man-machine relationship, moreover – it’s also beautiful. Taking a rudimentary Chess program, with each turn it exposes the potential moves the computer is considering through a gracefully sketched diagram overlaid onto the board. The experience reveals its real depth only in the play of a full game. As the machine gradually refines and rejects strategies, the narrative of the game thinking before you becomes more and more compelling. The developer has posted an example gallery which demonstrates this well, although this is well worth a few minutes of your quiet attention to actually play.