While the link posted is quite nice, I usually find it sufficient to simply ask what, exactly, is meant by "information" - then watch them hem and haw as they pour over the creationists sites they got this ridiculous argument from and come up with next to nothing.

Well I think the idea is that if nothing is added then things can't get more complex.

One of the most important outcomes of gene duplication is the origin of novel function. Although it seems improbable that an entirely new function could emerge in a duplicate gene, there are several examples. For instance, the eosinophil-derived neurotoxin (EDN) and eosinophil cationic protein (ECP) genes of humans were generated in the lineage of hominoids and Old World monkeys via gene duplication [42]. Both genes belong to the RNase A gene superfamily. After duplication, a novel antibacterial activity emerged in ECP. This activity is absent in human EDN and the EDN of New World monkeys, which represents the progenitor gene before duplication. More surprisingly, the antibacterial activity of ECP does not depend on the ribonuclease activity [43]. Molecular evolutionary analysis suggests that the new function is probably conferred by a large number of arginine substitutions that occurred in a short period after duplication [42]. ECP is toxic to bacteria because itmakes their cellmembranes porous; thepositively charged arginine residuesmight be important for establishing tight contact between the ECP and negatively charged bacterial cellmembranes in the pore-formation process [42]. In many cases, however, a related function, rather than an entirely new function, evolves after gene duplication.  One good example is the red- and green-sensitive opsin genes of humans, which were generated by gene duplication in hominoids and Old World monkeys [44]. After duplication, the two opsins have diverged in function, resulting in a 30-nm difference in the maximum absorption wavelength. This confers the sensitivity to a wide range of colors that humans and related primates have.

Amino acid sequence comparisons have been made between all of 25,193 human proteins with each of the others by using BLAST software (National Center for Biotechnology Information) and recording the results for regions that are significantly related in sequence, that is, have an expectation of <1 x 10–3. The results are presented for each amino acid as the number of identical or similar amino acids matched in these aligned regions. This approach avoids summing or dealing directly with the different regions of any one protein that are often related to different numbers and types of other proteins. The results are presented graphically for a sample of 140 proteins. Relationships are not observed for 26.5% of the 12,728,866 amino acids. The average number of related amino acids is 36.5 for the majority (73.5%) that show relationships. The median number of recognized relationships is ≈ 3 for all of the amino acids, and the maximum number is 718. The results demonstrate the overwhelming importance of gene regional duplication forming families of proteins with related domains and show the variety of the resulting patterns of relationship. The magnitude of the set of relationships leads to the conclusion that the principal process by which new gene functions arise has been by making use of preexisting genes.

The precise count of repeats is obviously underestimated because the genome sequence is not finished, but their density and other properties can be stated with reasonable confidence. Currently recognized SINEs, LINEs, LTR retroposons and DNA transposon copies comprise 13%, 20%, 8% and 3% of the sequence, respectively. We expect these densities to grow as more repeat families are recognized, among which will be lower copy number LTR elements and DNA transposons, and possibly high copy number ancient (highly diverged) repeats.

Analysis of the human genome revealed that some 45% of it consists of various kinds of transposable elements. Around 8% of the human DNA is derived from retrovirus-like elements. They originate from ancient retroviral infections or are relics of retroviral transposomal activity in the germ-line cells. Human endogenous retroviruses (HERVs) comprise a part of these elements. They have undergone substantial changes such as mutations of all kinds, deletions and insertions of other transposons, recombinations and mini- and micro-satellite expansion. This is why it is often difficult to identify individual retroviral genes and other retroviral DNA regions.

The retroviruses carry within their virion the enzyme reverse transcriptase, which enables them to make a DNA copy of their RNA.  This is a very important step in the life cycle of these viruses because the DNA copy can be integrated into the cell's DNA. (330)

Results of protein sequence comparison at open criterion show a very large number of relationships that have, up to now, gone unreported. The relationships suggest many ancient events of gene duplication. It is well known that gene duplication has been a major process in the evolution of genomes. A collection of human genes that have known functions have been examined for a history of gene duplications detected by means of amino acid sequence similarity by using BLASTp with an expectation of two or less (open criterion). Because the collection of genes in build 35 includes sets of transcript variants, all genes of known function were collected, and only the longest transcription variant was included, yielding a 13,298-member library called KGMV (for known genes maximum variant). When all lengths of matches are accepted, >97% of human genes show significant matches to each other. Many form matches with a large number of other different proteins, showing that most genes are made up from parts of many others as a result of ancient events of duplication. To support the use of the open criterion, all of the members of the KGMV library were twice replaced with random protein sequences of the same length and average composition, and all were compared with each other with BLASTp at expectation two or less. The set of matches averaged 0.35% of that observed for the KGMV set of proteins.