There seems to be alot of hype regarding epigenetics. Many of the grander claims seem to rest on thin evidence, and to the extent a phenomenon is real often given simpler explanations--cultural, genetic, or epigenetic--after everybody stops paying attention.

The evidence is actually quite dramatic. Birds that have difficulty cracking open the seeds available to them give birth to birds that have much thicker and strong beaks. Birds that have trouble extracting seeds from large flowers or other structures give birth to birds with longer and thinner beaks. Large changes to phenotype that used to be thought to evolve only slowly over long periods turn out to happen as quick changes based on environmental inputs.

Epigenetics is very mainstream now. "Big data" was hyped once too, but now MapReduce and stream processing are common and boring.

Just a data point:

mapreduce and stream processing were well published things long before Google popularized them, in the world of parallel computing at least back into the late 80s, when i started grad school doing just those things.

sadly i donated my old books to the library years ago or i'd provide a citation.

I think you missed a point. It is not about passing on a trait to your children but about passing on an idea to anyone else. In this context selective pressures apply as information spreads across individuals

The part the parent refers to IS about passing on a trait to children. It's not in the part of the article talking about mimetics yet.