In this lecture, I discuss two 'hard' problems in the numerical simulation of many-body systems. The first is the quantitative study of the slow (algebraic) decay of tagged-particle correlation functions in simple fluids. The second is the numerical study of phase equilibria in polymer systems.Using conventional simulation techniques, many aspects of these problems cannot be studied with existing (super)computers.However, if we consider `lattice-gas' versions of the same problems, then it turns out that a speed-up of 6-10 orders of magnitude can be achieved using a very simple algorithm. As a result, we can now test theoretical (mode-coupling) predictions for long-time tails with unprecedented accuracy. Similarly, certain phenomena in polymer physics, such as the polymer-induced clustering of colloids, that did thus far defy simulation, can now be observed and studied numerically.