if the plants have other ways to spend chakra on accelerating their own growth, that doesn't change my conclusion at all, so long as there's still an attainable point of overall diminishing returns - and I'm reasonably certain things would look very different if there weren't.
I agree that there likely is a point of diminishing returns. However, we know for a fact that increasing chakra absorption capacity to the point it's at now for this strain increased growth. It therefore follows that, at the level of chakra intake this plant has available, it has not yet hit diminishing returns - at least, not to the point that the effect from reduced chakra availability would be irrelevant.
The pulsed vs. continuous expenditure distinction is also not relevant, nor the exact ratio of chakra intake to storage capacity.
You described this chakra-harvest rune as operating by taking excess chakra above a certain level. If expenditure is pulsed, then the plant could reasonable be expected to sit at low chakra reserves during growth (since it would have to regularly activate the growth technique and drain its reserves) and so would dodge the effect of the rune. If, on the other hand, expenditure is continuous, there's no reason for the plant not to sit at high chakra capacity (in order to have chakra available for defence) and be subject to the rune's drain constantly - and probably respond by throttling chakra flow to growth techniques, since it would expect to get that chakra back relatively quickly and would want to have a reserve available for defence.
On the XP issue, I cannot reasonably be expected to rebut an argument (or accept, and derive new conclusions within, an alternative model) which you don't bother to spell out.
Fine. The level system in MfD is a system principally intended to model human skill growth (effectively, training of a neural network); as you reach higher levels in skills, you require more and more training to asymptotically approach perfection, and/or the number of different things you have to be good at increases, hence the linear increase in XP needed to buy the next level. This is, itself, an abstraction, and flawed in a number of ways - but it's not a terrible abstraction, it does at least reproduce the rough behaviour we should narratively see. The few non-neural skills (Physique, Chakra Reserves) can be explained by the increased difficulty of pushing further from homeostasis (humans are not adapted to have more muscle than they need day-to-day), and Doylistically by the QMs not wanting to model them differently. (Though this is one of the major flaws in the system - we've heard narratively of people having variably sized chakra reserves for genetic reasons, but there's no mechanical provision for this. Similarly, adults ought to be stronger than children, but there's not provision for that either. I'm inclined to suspect that physical skills for ninja are actually almost entirely chakra-based, in which case they can display this behaviour because chakra is weird.)
Plant growth does not work like this. Plants are not growing by training neural networks, they're growing by accumulating biomass. (Or rather, chakra plants may indeed have neural networks; but we're not talking about that kind of growth.) That accumulation of biomass is exponential, in that the more mass the plant has the more light it can absorb and the more carbon it can fix, and is extremely poorly modelled by the MfD level system. (I'm aware that the level system is likely nonlinear; but the point stands regardless.)
To put it another way: the MfD level system explicitly fails to account for developmental changes (example; a 20-year-old adult has the same Physique progression as a 13-year-old child) - whether because of Doylist abstraction or because those changes are insignificant in the face of chakra, it doesn't matter. In the chakra rice, nearly
every change is developmental, and so trying to model it with XP seems to me to be a poor choice.