> Even if the animal can position itself into a teeter-totter position with center of mass over the pivot/legs, it would still be using muscular energy to straighten up and extend, and then coming back down can hardly have been lossless

The idea is that they don’t position themselves into a teeter-totter position, they always are in that state (at least the species with very heavy tails). When they’re on their front feet, they’re essentially leaning forward. By bringing their tail down, they lean back into a rearing position.

That said I think most paleontologists think neck-feeding is far more likely than rearing because of the energy argument. There are relatively strong arguments for the latter along other axes but we don’t have any fossils locked in that position to prove it.

> it'd be a combination of again using muscles to come down in a controlled manner (and not destroy it's front joints!), and then a final plop down which would transfer kinetic energy into compressing the landing spot...

These dinosaurs were so heavy these were all problems already. Just supporting the mass of the skeleton required a bunch of special adaptations, some of which we see in elephants. The problem is that there aren’t many extant megafauna analogues and for many of these animals we’re not even sure how their metabolisms worked, let alone the intricacies of their musculoskeletal adaptations.

> all for a mouthful of leaves.

Vast majority of the giant necked herbivores didn’t chew at all to keep their heads light, their teeth designed instead to cut and strip entire branches at a time into their gut where they likely fermented. Regardless of the feeding method, they didn’t work a mouthful at a time.