I actually had a conversation about this earlier today, but overall I don't think I mind a whole lot.

Chasing the next popular bit of research 10 years in advance is a bit of a fool's errand, so the best you can do is follow along with fields that you think are personally interesting.

For a more serious answer, I do think that FPGAs are going to become much more applicable to general-purpose-computing soon. Most of the tooling and language support around them is pretty archaic, and tools like CλaSH [1] make development much more friendly.

[1] http://www.clash-lang.org

Look up MyHDL.