AI as Search

When I took my first AI class (1986?), we spent a large portion of the semester working on search.
When Deep Blue beat Kasporov, it used a simple 19 ply A* with pruning algorithm.
In 86, the feeling was that AI needed to have improved search algorithms to succeed.
I spoke with Marvin Minsky a few years ago, and he still felt that way.
One key to this is state space. You describe chess, most problems, and to some extent the world, by a series of states.
The starting state of a chess game, has all 32 pieces in their particular place.
You move from state to state by moving pieces (in a turn).
You can evaluate how good a state is.
So, the problem becomes searching for the best (or a good) state given your current state.