It generally matters little, if at all -- the optimizer should be able to use statistics to order the predicates optimally.
Far more important are statistics and indexing. Without an appropriate index, the query will have to scan the table or some index. If the first n columns in an index match search predicates, the query engine can seek or at least scan a subset. Without up-to-date statistics, the optimizer may make bad guesses on which index to use and how to use it.
An index on FirstName and LastName, with one of them being the first column in the index and the other being the second column in the index, would probably be able to directly seek the matching rows with either of these queries.
An index starting w/ FirstName (but not LastName) would be able to find all the "F"s, but would then still have to scan through the LastNames to find "L".
An index starting w/ LastName (but not FirstName) would be able to find all the "L"s, but would then still have to scan through the FirstNames to find "F".