One bad thing about Gua\spi numbers is that they lock in the Frege-Russell interpretation of natural numbers, in which 0 is the set of the null set, 1 is the set of all singletons, 2 is the set of all unordered pairs, and so on. These sets are infinite, which makes them hard to handle in practical implementations. It would be very awkward in Gua\spi to make use of the alternative Von Neumann interpretation, in which 0 is the empty set {}, 1 is {0, {0}}, 2 is {1, {1}}, and so on, because there is no straightforward way to construct sets by simple enumeration.
There are other possible interpretations: any will do as long as the Peano axioms are obeyed. Lojban numbers are abstractions, and as such don't lock in any of these interpretations.