In 2004, the genius Hawking admitted he had been wrong and conceded a bet he made in 1997 with a fellow scientist about black holes. To understand the bet, let's backpedal a little to understand what black holes are in the first place.
Stars are gigantic -- they have so much mass that their gravity is always incredibly strong. This is fine, as long as the star continues to burn its nuclear fuel, exerting this energy outward, thus counteracting gravity. However, once a massive enough star "dies" or burns out, gravity becomes the stronger force, and causes that big star to collapse on itself. This creates what scientists call a black hole.
The gravity is so powerful in this collapse that not even light can escape. However, Hawking proposed in 1975 that black holes are not really black. Rather, they radiate energy.
But, he said at the time, information is lost in the black hole that eventually evaporates. The problem was that this idea that information is lost conflicted with the rules of quantum mechanics, creating what Hawking called an "information paradox."
American theoretical physicist John Preskill disagreed with this conclusion that information is lost in black hole. In 1997, he made a bet with Hawking saying that information can escape from them, thus not breaking the laws of quantum mechanics.
Hawking is such a good sport that he can admit when he's wrong -- which he did in 2004. While giving a lecture at a scientific conference, he said that because black holes have more than one "topology," and when one measures all the information released from all topologies, information isn't lost [source: Rodgers].