A reader who goes by the handle "biologist," and happens to be a molecular geneticist by training, states more clearly what is probably close to my own position (though he is far more well informed) in the comments below. I think it's worth promoting:
As far as I can tell, the existence of epigenetic mechanisms doesn’t change anything that we *should* have already known about the social implications of genetics (i.e. what people care about). Quantitative genetic methods that estimate a substantial contribution of genetic variation to phenotypic variation do not now and has never told us anything about actual or counterfactual causal mechanisms involved. They have also never told us much about development other than what we already knew must be true — there will be genes involved in some way. Nothing we’ve learned in the last 30 years about molecular biology makes any difference at a general level to those conclusions. What it mostly does is make clearer that the causal mechanisms behind phenotypic variation in complex traits are probably themselves really really complex. As soon as you realize that complex traits have non-Mendelian inheritance patterns — something that’s been abundantly clear for many many decades — everything else follows and epigenetics only adds new dimensions to the causal mechanisms that might be involved. Whether a trait is amenable to manipulation (and at what stages of development) is an interesting and very challenging question, but there’s no revolution in our understanding of biology involved in asking it. The only way to see a revolution is to ignore all of the incremental changes in understanding that have happened between decades.
Just to be clear, this is not a very mature sounding 12 year old. The commenter above is a biologist with whom I am personally familiar and whose opinion on this topic I value because not only do they grasp molecular biology in its fine-grained details, but they are very familiar with quantitative and behavior genetics (a rare combination). I can probably transfer some of the same general cautions about epigenetics that I brought up with Jim Manzi in relation to epistasis several years back. The great thing about science is that this likely won't be a debate 10-20 years from now. If you have an equation of the form: A[genetics] + B[epigenetics] + C[environment] → Outcome The scalars A, B, and C will be known with more precision as science progresses. Or more accurately, their values will be known for the range of outcomes which we find of interest. Our current surfeit of commentary is a function of mystery and uncertainty.