It can only be solved in that a correct answer can be given; there are some Turing machines that do not halt for which there is no proof that they do not halt.
Yes, this is correct, but the halting problem can still be solved with a simple algorithm for each such machine, and even for all of them at once :-) But okay, we understand each other.
A monthly thread for posting rationality-related quotes you've seen recently (or had stored in your quotesfile for ages).