They show that a measurement of the real space correlation function might enable a distinction to be made between “causal” sources of CMB B-modes (here the authors take “causal” to mean non-inflationary sources, so cosmic strings, phase transitions, etc fall into this category) and inflationary sources.

in simple models (there’s that word again)

If you don’t like it, then YOU can go solve third order non-linear differential equations!

Secondly, without particle production, its generally true that the amplitude of gravity waves is set by the Hubble scale during inflation (its just due to the de Sitter temperature T=H/(2*pi) causing fluctuations in the metric, not just in “simple models” as Sean is claiming).

Thirdly, the basic idea that gravity waves from the early universe may not tell us the details about inflation was made much earlier in a much more dramatic example in http://arxiv.org/abs/0712.0778, where Smith, Krauss, and Mathur point out that even a global phase transition can reproduce the same kind of amplitude and spectrum of gravity waves. This is a much more startling conclusion than the Cook and Sorbo result or the above result, since it can occur after inflation or even if there is no inflation at all!

True, but much better than having no practically testable predictions at all, eh?

And until “compelling models” can make practically testable predictions that cannot be “adjusted” as needed, or conveniently ignored when problematic, they are just so much hot air.

J.~L.~Cook, L.~Sorbo,
“Particle production during inflation and gravitational waves detectable by ground-based interferometers,”
[arXiv:1109.0022 [astro-ph.CO]].