Thermosetting plastics of the epoxy family have two modes of polymerisation. The fundamental mode forms large elongated molecules, with (relatively) rigid zig-zag shape, and a few cross-links to adjacent molecules. The GTT is, crudely speaking, when the "free" energy level exceeds the cross-link energy sum for each molecule - wtf? well, hotter = more amplitude of vibration of the atoms, and when they are stretching the intermolecular bonds too far, the bonds can "snap" across to the nearest adjacent attachment site. So, them there rigid zigzag molecules start to slide past each other.
Now, the secondary mode of polymerisation of thermoplastics, involves a lot of cross links forming close together; if excited by enough heat - as in staged elevated temperature curing - the zig-zag molecules can actually get a bit shorter, as the intra-molecular bonds are reformed into inter-molecular bonds.
The more inter-molecular bonds there are, the higher the GTT becomes. So, a room-temperature cured epoxy has a low GTT as cured; running it through temperature cycles - such as day and night through half a year - gradually elevates the GTT.
So why not autoclace the crap out of it? Because the maximum rupture strength of the epoxy matrix is reached at less than complete cross-polymerisation; AND, critically, the maximum fatigue endurance is reached at rather less cross-polymerisation than the maximum rupture strength. The Jabiru "system" - which includes chemicals, application temperatures and times, and cure cycles - achieves outstanding fatigue resistance, at adequate rupture strengths, using low-overhead production techniques (otherwise nobody could build LSA-E kits!).
Rotor blades are rather specialised, in that the centrifugally induced stresses dominate; their fatigue resistance has to be optimised only for a spanwise mean-plus-cyclic situation, and no triaxial restraint exists in the loaded area, unless the designer has gone to truly extraordinary lengths to screw up!
I have had the - pleasure? - of designing a repair scheme for a set of doors on a certain model of FAR-23 composite piston single, which at 1100 hrs TT (then the oldest in the world fleet!) had developed fatigue cracks above both door-mounted windows, from slipstream tip vortex impingement. They were autoclaved carbon fibre doors, strong enough to kill oxen, but they couldn't handle fatigue.
Jabiru put loads of work into structural integrity / composite fatigue research, and when it comes to resilience, they are the (composite) light aeroplane gold standard. Note that I hold no shares in Jabiru, do not own a Jabiru, and have no present intention to get one; but I have (somewhat grudgingly) acquired a high respect for Rod's approach to composite structure,