WTF Bayesian Networks

So it would seem every where you turn in the defense contractor world, every one is running around with their heads cut off wanting Machine Learning. Haven't these people seen Terminator 1-100? I even read scientist have developed liquid that behaves similar to the T-1000. For my use, I am applying Bayes net to some AI for unmanned ground vehicles. Also apply predator algorithm so the robot could know what the hell its looking at. I have yet to try out any of these ideas. Like many people that publish white papers, I am all talk and no doing thus far. Objection! I must do before I talk! So I have been trying to apply Bayes net to my particular problem domain and it is a bitch! I have what they call purple link syndrome. Where all the links from google have been looked at already. Funny thing is i feel much further away from a solution then when I began. The more I read the more I am reassured that its hard to apply Bayesian Reasoning. I have even cracked open an old dusty book on AI. I have a feeling, even if I fully understand the concept of Bayesian Reasoning, like expert systems, when I try to describe it to someone, they go crossed eyed. This makes me weary of either I do not know how to teach people or they just refuse the concept. So I dumb it down. Then they want more detail. So I detail it up. They go cross eyed. So I guess the true art here is finding the nice sweet spot inbetween where you are not to technical and not to vague.