Hill climbing is an optimization method used in artificial intelligence which involves iteratively making small changes to a solution and evaluating whether it improves or worsens the solution. The process of hill climbing works by trying different solutions around a current solution, then selecting the solution that will give the greatest benefit with the smallest amount of effort. It is analogous to traversing a hill with slopes in all directions, where the goal is to reach the highest point. Hill climbing is a type of local search and works best when the solutions are comparably small and finding an optimal solution isn't critical.
See also: adjacent possible, artificial intelligence, feedback loop, game theory, low-hanging fruit