Inference is dynamic: It changes over time; it exists in a changing environment (the world is constantly changing, the actors who draw inferences change); inference suffers the slings and arrows of changing preferences, emotions, desires, passions, etc.; inference changes when subjective judgments change and when generalizations or beliefs about the world change; it often changes when the evidence changes, which evidence always does (to some degree); inference changes when standards of judgment (e.g., burdens of persuasion, or what decision theorists call "decision rules") change; and inference changes for yet other reasons I can't think of at the moment.
Fact investigation is also hard in part because, as someone, said, "Predictions are hard, particularly about the future." An investigator, who always suffers from some ignorance, must make decisions about what to investigate and what lines of investigation are worth pursuing. Such prophecies about future evidentiary rewards are intrinsically hard to make -- given the investigator's ignorance of the evidence that will ultimately be found (or not found). Cf. "value of information" theory.