I work primarily in extremal and probabilistic combinatorics. Most problems of an extremal nature involve a set (_nite) with certain constraints (which are of a combinatorial nature) and one is then interested in how large/how small such a family could be. In classical problems of extremal combinatorics, the extremal families also admit short (low complexity) descriptions. However, the more interesting ones are those where the class of extremal families is very large, without any seeming sense of structure, other than the fact that they are extremal families of a certain combinatorial extremal problem. Interesting instances of these occur in the study of graph coloring problems where one imposes di_erent kinds of restrictions on the coloring.