In a sensor network monitoring natural environment, the readings of sensor nodes show high temporal and spatial correlation. This stems from the fact that most characteristics of natural environment (chemical concentration, temperature, humidity, pollution, etc) do not change abruptly in space and time. From a sensor's perspective then, the environmental phenomena are quite predictable, and hence, "boring". In this paper, we introduce the novel idea of buddies to exploit temporal correlation. We propose a simple Buddy protocol to implement this idea. We analyze the performance of the protocol and identify factors that influence it. We discuss research challenges that would pave the way for more efficient protocols based on this idea.