I actually spent quite a while in grad school thinking about the last parsec problem, and although I'm not in the field anymore I still think about it from time

I actually spent quite a while in grad school thinking about the last parsec problem, and although I'm not in the field anymore I still think about it from time to time. (My thesis was on gravitational dynamics.) My perception of the field (which is now about a decade out of date, so take it with a grain of salt), is that there is quite a bit of skepticism about invoking exotic physics to solve the last parsec problem. Galaxies are generally pretty messy places, and the centers of galaxies are especially messy, so it's hard to know if you've correctly modeled all the relevant physics. A lot of astronomers aren't convinced that there really is a last parsec problem. The main "standard" approach to solve the last parsec problem is from scattering stars (which the article mentions). Basically, every now and then stars from the galaxy wander close to the orbit of the black hole binary and then get slingshotted out of the system. This removes energy from the orbit, and causes the black hole binary to shrink. The problem with this approach if you do a naive calculation is that the stars have to come from a particular set of directions, called the "loss cone" in the jargon. And since the orbits of stars in galaxies are probably fairly static, once a star gets kicked out of the loss cone, it doesn't come back. So over time the loss cone empties and the black hole orbit stops shrinking. The question is, does the orbit shrink far enough before the loss cone empties, and the answer to this question has generally been "no." The way around this is to question how static the orbits of stars in galaxies really are. One of the more important papers on the topic found that if an elliptical galaxy is sufficiently triaxial (that is, sufficiently non-spherical), then interactions between stars in the galaxy can repopulate the loss cone and cause the orbit to keep shrinking. But as I vaguely recall, not everyone was convinced by that result. I personally have had some ideas that galactic tides might contribute, especially right after the merger before all the orbits have had time to thermally relax. But I'm not in the field anymore and haven't really had time to really model this idea and see if it would work.