Since there are
4 3 = 64 residue triples formable from the alphabet {A,G,U,C}, there is obviously a good deal of redundancy, and because of this there is a loss of information. At best, not counting the ends, the information in an amino acid gene word of length 4 is
log2(204) =

log2(160000)

= 17.3 bits,

and this is only true if each of the letters occurs with equal probability (p=1/20). Which they do not. There are two reasons for this: some amino acids are coded for by more triples than others (click a few times on the button and compare the number of occurances of the letter "R" to the letter "M"); and some amino acids are biologically more important than others - prefered by Nature.