"Not breakable in fewer than O(2^(n/2)) operations" is an imprecise security claim; "challenging" means nothing. "Challenging" may not even mean "computationally hard"---breaking a truncated LCG is arguably quite challenging if you are unfamiliar with lattice reduction.
Do you lock your front door? Is your lock unpickable, or robust to attacks from national superpowers? What formal security guarantees does it give you?
Surely you must accept that there is some scope, somewhere, for better-than-nothing security, some amount of protection that is not robust against arbitrarily skilled adversaries, but nonetheless makes it harder to break your things. So either you need to show that it is not useful in this case specifically, or you need to show that this particular formulation in untenable. Arguing about words does not get us closer to an answer; I have shown a specific threat and a specific mitigation, either it helps or it does not.
I will continue to mark the security of PCG as "unwilling to quantify", which is strictly below "better than nothing". Debating the virtues of the latter is therefore pointless.
Nowhere does O'Neil discourage expert analysis of PCG generators. She does lack the ability to do direct research to the standards of the industry, but discusses related work and attacks in the paper, and has encouraged further research. I don't see how you can see the evidence presented in the paper and claim there is zero practical difference between it and the examples I listed, even just talking about today's adversaries.
