Getting conditional probability is more difficult than joint probability?

Question

While I am studying Variational Inference, I am told that

"In Graphical Model, to get the joint probability, $P(X_1, X_2, \dots, X_n)$, is easy but it is difficult to get the conditional probability $P(X_1|X_2, \dots)$ and variational inference is useful for this case."

Can I have some intuitive explanation about the sentence?