They're right about entanglement of many organizations with MS Office-based applications. Their judgement regarding strong copyleft being a barrier to replacing that tangle, I have no idea, would be curious to read more about it.
LO has been requiring contributions under MPL&LGPLv3+ from the beginning (or nearly so) of the project, and re-basing has been underway for a long time. The "switch" to MPL2 is long-planned and unsurprising. But ignoring all that, it would've been curious had LO instead gone GPLv3+, making for a tiny bit more of a natural licensing experiment relative to AOO (tiny bit as non-license differences probably dominate), and AOO would've been available for proprietary vendors and integrators in any case.