Rob Knop has a nice post on how simulations can never be "better" than experiments (er, in the sciences). For people like me whose professional work is centred on simulations of physical systems, this is a mantra that's worth repeating many times a day, just to keep our over-enthusiasm in check. While there are tons of neat (and advantageous) things about simulations (and many, many good reasons why simulations are important and worth doing), the meme that "simulations might (somehow) be 'better' than experiments" is not one of them.
My favourite quote about simulations:
No one gets wet in a simulated thunderstorm.
-- Cosma Shalizi