As sequences of protein-coding genes began to accumulate in the 1980s, it became apparent that different synonymous codons were used preferentially in different species. The phenomenon became known as codon bias. By 1990 it was known that the frequency of codon usage was correlated with tRNA abundance. As a general rule, there is a different tRNA for each codon and if multiple codons exist for a given amino acid then insertion of that amino acid into protein will depend on different tRNAs carrying the same amino acid.
Highly expressed genes are enriched for those codons with abundant tRNAs. This maximizes the speed of translation. Most of the first genes that were sequenced were highly expressed and that's why codon bias showed up so early. As more and more genes were sequenced we often saw that there was no preferential use of the "biased" codons.
There may be examples of genes that selected for rare codons in order to slow down translation and reduce the amount of protein being produced but there aren't many examples.
The idea that synonymous codons aren't always neutral has been around for decades but every now and then some science journalist discovers it for the first time [A Breakthrough in Gene Expression?] [Silent Mutations and Neutral Theory]. And it doesn't stop the Intelligent Design Creationists from claiming that it somehow disproves evolution and confirms a prediction by IDiots [Another stupid "prediction" by Intelligent Design Creationists].
Now they're at it again. The latest paper shows that by slowing down translation, the mRNA becomes less stable (duh!). Thus, rare codons play a role in regulating gene expression. This leads Casey Luskin to announce: Another Successful Prediction of Intelligent Design: Cell Paper Reports Functions for Synonymous Codons.
ID, on the other hand, predicts that we should find deeper and deeper layers of function in biology. That's why functions for synonymous codons represent a successful prediction for ID, and a big problem for Darwinian evolution.I don't recall that ID predicted codon bias back in the early 1980s. Do you?
And I don't recall a lot of angst at the time because this was a big problem for Darwinian evolution.