Why will we never be able write down all automorphic representations "explicitly"?

Question

In a recent article (p. 7), J. Arthur writes

In fact, it is pretty clear that we will never be able to write down all automorphic representations explicitly.

I have often heard similar remarks made in talks, but I have never pursued an expert for an answer. I guess there are two sub-questions that I would like answered:

1. What does it mean to write down an automorphic representations "explicitly"?
2. Why is such a thing impossible?