In 1913 French mathematician Émile Borel proposed the infinite monkey theorem: A monkey hitting keys at random on a typewriter keyboard for an infinite amount of time will almost surely type a given text, such as the complete works of William Shakespeare. Now while there is a certain resemblance to the internet :), collective problem solving on the internet works in fact much better than randomly hitting keys. While an individual player might be solving a problem in a game by trial and error, he can then exchange the results with others via various internet platforms, either in written or video form. With many people collaborating to solve a game problem by exchanging what works and doesn't work, the speed of problem solving becomes a lot faster than infinite.
That does have repercussions on game design. If you have a puzzle in your game, you not only need to think about how different players might have different skills to solve your puzzle, but also take into account that some will tend to try to solve it alone, while others will look up the solution before even having seen the puzzle. But even more striking is the effect of collaborative problem solving on game balance. If your game has an optimum path, players will find it through collaboration.
I was thinking about it when reading various blog posts about Hearthstone. I played a lot of Magic the Gathering in paper and online form in its time, and Magic the Gathering was reasonably well balanced: While at any given expansion a certain deck might be found to be "the best" of a specific play style, there was always a stone-paper-scissors meta-game in which one play style would do well against another, but lose against a third. And I am not certain that Hearthstone has that stone-paper-scissors meta-game, because the little I have watched of it so far shows it to be a much simpler game than Magic. It is also a CCG [collectable card game] not a TCG [trading card game].
Now of course it is hard to judge from excited blog posts about one deck being completely overpowered to judge the actual state of the game. But to me it seems that Hearthstone is more likely to "get solved", that is an optimum deck found for any given set of cards. While every expansion or nerf then changes that optimum, once the fundamentals are understood, the collective problem solving will find the new optimum while the changes are still on the test server. It is quite likely that the release version of Hearthstone will "get solved" before even leaving the beta. That will affect the game's longevity.
At the very least some cards will be found to be better than other cards. So while the game pretends to have a certain number of different cards, those in the know will work with a much smaller card pool. And those who absolutely want to win will have to spend more money to get sufficient numbers of that smaller, better card pool, because suboptimal cards keep popping up in the booster packs they open. It is not possible to have 600+ cards all with the same power, and I hope that too much tinkering does not spoil the final game.