The ‘genomic shock’ hypothesis posits that unusual challenges to genome integrity such as whole genome duplication (WGD) may induce chaotic genome restructuring. Decades of research on polyploid genomes have revealed that this is often, but not always the case. While some polyploids show major chromosomal rearrangements and de-repression of transposable elements (TEs) in the immediate aftermath of WGD, others do not. Nonetheless, all polyploids show gradual diploidization over evolutionary time. To evaluate these hypotheses, we produced a chromosome-scale reference genome for the natural allotetraploid grass Brachypodium hybridum, accession ‘Bhyb26′. We compared two independently-derived accessions of B. hybridum and their deeply diverged diploid progenitor species B. stacei and B. distachyon. The two B. hybridum lineages provide a natural timecourse in genome evolution because one formed 1.4 million years ago, and the other formed 140 thousand years ago. The genome of the older lineage reveals signs of gradual post-WGD genome evolution including minor gene loss and genome rearrangement that are missing from the younger lineage. In neither B. hybridum lineage do we find signs of homeologous recombination or pronounced TE activation, though we find evidence supporting steady post-WGD TE activity in the older lineage. Gene loss in the older lineage was slightly biased toward one subgenome, but genome dominance was not observed at the transcriptomic level. We propose that relaxed selection, rather than an abrupt genomic shock, drives evolutionary novelty in B. hybridum, and that the progenitor species’ similarity in TE load may account for the subtlety of the observed genome dominance.