The "occasional minor pains" example is problematic because it brings in the question of aggregation too - and respective problems are not specific to NU. If NUs have to claim that sufficiently many minor pains are worse than torture, then that holds for CUs too. So the crucial issue is whether the non-existence of pleasure poses any problem or not, and whether the idea of pleasure "outweighing" pain that occurs elsewhere in space-time makes sense or not.

It's clear what's problematic about a decision to turn rocks into suffering - it's a problem for the resulting consciousness-moments. On the other hand, it's not clear at all what should be problematic about a decision not to turn rocks into happiness. In fact, if you do away with the idea that non-existence poses a problem, then the NU implications are perfectly intuitive.

Regarding Ord's intuitive counterexamples: It's unclear what their epistemic value is; and if there is any to them, CU seems to be subject to counterexamples that many would deem even worse. How many people would go along with the claim that perfect altruists would torture any finite number of people if that would turn a sufficient number of rocks into "muzak and potatoes" (cf. Ord) consciousness-seconds? As for "making everyone worse off": Take a finite population of people experiencing superpleasure only; now torture them all; add any finite number of tortured people; and add a sufficiently large number of people with lives barely worth living (i.e.: one more pinprick and non-existence would be better). - Done. And this makes you a good altruist according to CU.