I think Pacatus you have raised the real issue underlying the question - which includes the relationship and use of both direct evidence and "circumstantial" or "indirect" evidence.
Epicurus has no direct observational evidence of atoms or void, but he found strong circumstantial evidence from the observation of other things which he could observe directly. We therefore logically do not require direct evidence on all issues, but we require that any theories be consistent with what evidence does exist.
So when we use words like speculation and evidence we have to be very clear what we are talking about.
As I see it there is a LOT of circumstantial evidence from all else that we observe that "nothing comes from nothing..". We do not have direct evidence of what came before the "big bang" (which itself we infer from circumstantial evidence as we did not see it occur).
It is my personal position that the overwhelming evidence of nothing coming from nothing in all other observations is not invalidated as a good working model of the universe as a whole by the argument that "we have no direct evidence of what came before the big bang."
These are complicated epistemological issues that cannot be resolved purely by demanding "direct evidence* from our own observation. In order not to be accused of blind faith ourselves we have to have an understanding of our own rules of reasoning and waiting and opinion making as in the PDs in the early 20's