Deep Fritz has defeated Kramnik in their match, four games to two.
It is now clear that we are inferior to computers in chess, and this brings a new era where chess is more like a video game now, where the software corrects our mistakes by defeating us, and the game can be "solved" beyond what we were previously capable of.
In the "dinosaur" days of chess, grandmasters had to guess if their moves were sound. If something looked good or logical, it was played, and even after many games, we couldn't be certain if the moves were the best. Now, however, computers can solve highly complex positions and shed light on many more. What's most interesting is that the computers are able to win by playing openings that give it mere equality, at which point their tactical strength kicks in as they suffocate their helpless GM opponents in the middlegame. Computers have also verified that many "fringe" openings (such as the Latvian or the Center Counter) are much richer in possibilities than we humans ever dreamed possible.
Chess is in its golden age right now: computers are strong enough to help the average person understand the game and watch high-level play, but not so strong that we can't outplay them from time to time, or that they can't get stumped by complex positions. At some point, the game will be reduced to a complex algorithm (we're already up to six-piece tablebases), and the winner will be the one who memorizes existing theory. There will be new theory developed along the way, but the focus in chess has definitely shifted from the creativity required to develop new theory instead to the ability to master what is already known.