I guess it depends on the computational technique one is using, but for molecular simulations i don't think we are solving anything. We are evolving the system according to some law, and then take averages over the explored phase space, which is qualitatively what also happens in experiments.

The former, in my opinion. They used to be called "computer experiments" for a reason!

Jokes on you, I only submit a paper when I start hating it, not when I think it's good😁

P.S. I signed up and am definitely looking forward to trying it!