and I talked with another seismologist yesterday with the same opinion, we're not sure what he means.

Especially now with GPS, we can watch elastic strain build between earthquakes, and measure its release in earthquakes. As I understand it, that is classic elastic rebound theory in action.

He talks about whether strain build-up or strength degradation is more influential in the timing of earthquakes. That is practically impossible to say. We watch the strain build, but who can measure the strength of a fault before it has begun to fail?

I don't see how fresh rock breakage vs pre-existing faults matters in this stress vs strength discussion.

I'm guessing he means the now discredited idea of uniform recurrence intervals between earthquakes, which might have worked if only stress build-up and a constant failure stress were in play, doesn't work. This would be a plausible argument, and the fact I can't figure out the article would then impugn my reading comprehension.